Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Number crunching :
Prime Meridian Day Challenge
| Author |
Message |
|---|
 Michael Gutierrez Volunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
                 
|
|
Welcome to the Prime Meridian Day Challenge
The eighth challenge of the 2022 Series will be a 10-day challenge celebrating the 1884 international agreement to standardize the official 0° longitude -- AKA the Prime Meridian! The challenge will be offered on the CUL-LLR and WOO-LLR applications, beginning 01 November 05:00 UTC and ending 11 November 05:00 UTC.
From nationaltoday.com:
Prime Meridian Day is marked on November 1 each year. In 1884, the prime meridian was defined by the location of the large Airy Transit Circle telescope at the Royal Observatory in Greenwich, England. Sir George Biddell Airy — the seventh Astronomer Royal — designed the legendary telescope in 1850. The Airy Transit Circle’s crosshairs defined 0° longitude. Today, we celebrate the prime meridian and its applications, including dividing the Earth into the Eastern and Western Hemispheres and setting up the global time zones.
Not gonna lie, I was hoping there would be some classification of prime number that had something involving "middle" or "meridian" in the name, but I came up blank. Let me know your nerdiest, most contrived connections to prime numbers you can make with this theme!
To participate in the Challenge, please select only the Cullen Prime Search (LLR) and Woodall Prime Search (LLR) projects in your PrimeGrid preferences section.
Note: The deadline for these WUs is longer than 10 days, so make sure your computer is able to return the WUs within the challenge time-frame. Only tasks issued AFTER the start time and returned BEFORE the finish time will be counted.
Note on LLR2 tasks: LLR2 has eliminated the need for a full doublecheck task on each workunit, but has replaced it with a short verification task. Expect to receive a few tasks about 1% of normal length.
Application builds are available for Linux 32 and 64 bit, Windows 32 and 64 bit and MacIntel. Intel and recent AMD CPUs with FMA3 capabilities (Haswell or better for Intel, Zen-2 or better for AMD) will have a very large advantage, and Intel CPUs with dual AVX-512 (certain recent Intel Skylake-X and Xeon CPUs) will be the fastest.
Note that LLR is running the latest AVX-512 version of LLR which takes full advantage of the features of these newer CPUs. It's faster than the previous LLR app and draws more power and produces more heat, especially if they're highly overclocked. If you have certain recent Intel Skylake-X and Xeon CPUs, especially if it's overclocked or has overclocked memory, and haven't run the new AVX-512 LLR before, we strongly suggest running it before the challenge while you are monitoring the temperatures.
Multi-threading is supported and IS recommended for CUL-LLR. (Tasks on one CPU core take about 4 days on average.)
Multi-threading is supported and IS recommended for WOO-LLR. (Tasks on one CPU core take about 7 days on average.)
Those looking to maximize their computer's performance during this challenge, or when running LLR in general, may find this information useful.
- Your mileage may vary. Before the challenge starts, take some time and experiment and see what works best on your computer.
- If you have a CPU with hyperthreading or SMT, either turn off this feature in the BIOS, or set BOINC to use 50% of the processors.
- If you're using a GPU for other tasks, it may be beneficial to leave hyperthreading on in the BIOS and instead tell BOINC to use 50% of the CPU's. This will allow one of the hyperthreads to service the GPU.
- The new multi-threading system is now live. Click here to set the maximum number of threads. This will allow you to select multi-threading from the project preferences web page. No more app_config.xml. It works like this:
- In the preferences selection, there are selections for "max jobs" and "max cpus", similar to the settings in app_config.
- Unlike app_config, these two settings apply to ALL apps. You can't chose 1 thread for SGS and 4 for SoB. When you change apps, you need to change your multithreading settings if you want to run a different number of threads.
- There will be individual settings for each venue (location).
- This will eliminate the problem of BOINC downloading 1 task for every core.
- The hyperthreading control isn't possible at this time.
- The "max cpus" control will only apply to LLR apps. The "max jobs" control applies to all apps.
- If you want to continue to use app_config.xml for LLR tasks, you need to change it if you want it to work. Please see this message for more information.
- Some people have observed that when using multithreaded LLR, hyperthreading is actually beneficial. We encourage you to experiment and see what works best for you.
Time zone converter:
The World Clock - Time Zone Converter
NOTE: The countdown clock on the front page uses the host computer time. Therefore, if your computer time is off, so will the countdown clock. For precise timing, use the UTC Time in the data section at the very top, above the countdown clock.
Scoring Information
Scores will be kept for individuals and teams. Only tasks issued AFTER 01 November 05:00 UTC and received BEFORE 11 November 05:00 UTC will be considered for credit. We will be using the same scoring method as we currently use for BOINC credits. A quorum of 2 is NOT needed to award Challenge score - i.e. no double checker. Therefore, each returned result will earn a Challenge score. Please note that if the result is eventually declared invalid, the score will be removed.
At the Conclusion of the Challenge
We kindly ask users "moving on" to ABORT their tasks instead of DETACHING, RESETTING, or PAUSING.
ABORTING tasks allows them to be recycled immediately; thus a much faster "clean up" to the end of a Challenge. DETACHING, RESETTING, and PAUSING tasks causes them to remain in limbo until they EXPIRE. Therefore, we must wait until tasks expire to send them out to be completed.
Please consider either completing what's in the queue or ABORTING them. Thank you. :)
About the Cullen Prime Search
A Cullen number in the form of n*2^n+1. Fr. James Cullen first studied these numbers in 1905 because they have interesting divisibility properties. It is conjectured that there are infinitely many n values that will yield Cullen primes, though it has been shown that almost all are composite.
Further reading:
About Woodall Prime Search
Woodall Numbers (sometimes called Cullen numbers 'of the second kind') are positive integers of the form Wn = n*2^n-1, where n is also a positive integer greater than zero. Woodall numbers that are prime are called Woodall primes (or Cullen primes of the second kind).
Further Reading:
What is LLR?
The Lucas-Lehmer-Riesel (LLR) test is a primality test for numbers of the form N = k*2^n − 1, with 2^n > k. Also, LLR is a program developed by Jean Penne that can run the LLR-tests. It includes the Proth test to perform +1 tests and PRP to test non base 2 numbers. See also:
What is LLR2?
LLR2 is an improvement to the LLR application developed by our very own Pavel Atnashev and stream. It utilizes Gerbicz checks to enable the Fast DoubleCheck feature, which will nearly double the speed of PrimeGrid's progress on the projects it's applied to. For more information, see this forum post.
____________
Time flies like an arrow. Fruit flies like a banana.
| |
|
|
|
Not gonna lie, I was hoping there would be some classification of prime number that had something involving "middle" or "meridian" in the name, but I came up blank. Let me know your nerdiest, most contrived connections to prime numbers you can make with this theme!
Searching for midday prime brought up some Indian TV news programs.
A prime p (other than the first one) has two neighbor primes, the preceding prime and the next prime. Either could be closer to p. But in the case where both neighbors are equally close, can we not say p is midday or noon? The usual term for such a p is balanced prime.
/JeppeSN | |
|
|
|
Not gonna lie, I was hoping there would be some classification of prime number that had something involving "middle" or "meridian" in the name, but I came up blank. Let me know your nerdiest, most contrived connections to prime numbers you can make with this theme!
Searching for midday prime brought up some Indian TV news programs.
A prime p (other than the first one) has two neighbor primes, the preceding prime and the next prime. Either could be closer to p. But in the case where both neighbors are equally close, can we not say p is midday or noon? The usual term for such a p is balanced prime.
/JeppeSN
Now, if such a number were also happy, it would be a happy balanced prime!
The smallest of which is 263.
____________
Eating more cheese on Thursdays. | |
|
|
|
|
In reply to both ^ and https://www.primegrid.com/forum_thread.php?id=10024&nowrap=true#157564
How about Meridian of Time Primes? Kinda has a nice ring to it, haha. Maybe some kind of megaprime that has something to do with the middle, like perhaps primes associated wit hthe Riemann Zeta Function and Siegel zeroes. Who knows?
____________
| |
|
|
|
|
Question,
I have an 15 12600 core with 6 cores and 4 efficiency cores,, that equals 10
If i put mt Use CPUs setting to 65 % it runs 10 work units.
I have 32 GB of ram if it matters.
Would this be the Optimum setup for the Cullen/Woodall challenge ? | |
|
Nick
Send message
Joined: 11 Jul 11
Posts: 2298
ID: 105020
Credit: 8,425,326,377
RAC: 6,766,236
                            
|
Question,
I have an 15 12600 core with 6 cores and 4 efficiency cores,, that equals 10
If i put mt Use CPUs setting to 65 % it runs 10 work units.
I have 32 GB of ram if it matters.
Would this be the Optimum setup for the Cullen/Woodall challenge ?
I think we are in unknown territory.
I think trying 1 or 2 tasks at a time.
If you can set affinity - what cores are used - 2 tasks may be good each with 3 good and 2 efficiency. If you can set affinity this way, I recommend that. I also have no idea, but it is going to be 1 or 2 tasks to run in cache. | |
|
|
|
Question,
I have an 15 12600 core with 6 cores and 4 efficiency cores,, that equals 10
If i put mt Use CPUs setting to 65 % it runs 10 work units.
I have 32 GB of ram if it matters.
Would this be the Optimum setup for the Cullen/Woodall challenge ?
Definitely in some new territory. We don't have a lot of data available about multithreading performance of mixed P+E cores.
Theoretically, you'll want to use all 10 threads per WU, as task size is right about the same as your L3 size.
Now, also something to test, if you want, is to find out (in the BIOS) if you are able to disable the E cores and enable AVX-512. Running with 6 threads/task, this will give you the best performance, but it has some dependencies on whether or not it will work: Your chip would need to be from the early batches that didn't have the AVX512 fused off and your BIOS would have to be old enough still have the settings.
____________
Eating more cheese on Thursdays. | |
|
|
|
|
I'm sitting here scratching my head on how on the leader board, we have people with completed tasks already. Then it hit me. I'm guessing its proof of work tasks.
By the luck of draw, it looks I didnt get any proof of work tasks.
| |
|
|
|
I'm sitting here scratching my head on how on the leader board, we have people with completed tasks already. Then it hit me. I'm guessing its proof of work tasks.
By the luck of draw, it looks I didnt get any proof of work tasks.
Looks like you posted at 10:27 almost 5 hours after it started.
My fastest woo tasks are done in 4.4 hours (16,000 and 17,000 seconds) using MT. | |
|
|
|
Looks like you posted at 10:27 almost 5 hours after it started.
My fastest woo tasks are done in 4.4 hours (16,000 and 17,000 seconds) using MT.
Thanks....
I switched to multi threading but I'm still at 13 hours..
ubuntu 20.04 using 8 threads on i11700k..
| |
|
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference! | |
|
|
|
|
Hello!
It seems I have a big problem with optimum settings.
I divided all my PCs at 5 different settings according to the quantity of CPU threads.
But the times are 24+ hours almost everywhere.
For example:
2 x Xeon 2470 (8c/16t)@2.6 are running Cullen with 16 threads per task. It took 23 hours to reach 94% of tasks. No AVX
2 x Xeon 2698v3 (16t/32c)@2.9 with 16 threads per task (AVX 2.0). 22 hours to reach 90%.
Ryzen 5900X (12c/24t)@4.2 with 12 threads per task 14 hours to reach 90%/
Everywhere HT is ON. CPU is fully loaded and my BOINC settings are set to use 100% of CPU cores and 100% of CPU time.
Is it normal? How can I improve my CPU output?
Thanks! And sorry for my bad English.
P.S. I've read about to set CPU cores to 50% in case when HT is ON. But when I set to use 50% of CPU cores my CPU load drops. What is the priority.
Set settings of CPU 50% and then download tasks or download tasks and set CPU to 50%
____________
 | |
|
Dave
Send message
Joined: 13 Feb 12
Posts: 3209
ID: 130544
Credit: 2,288,005,894
RAC: 763,651
                        
|
|
Try 50% CPUs not 100. Could be overloading the cache. | |
|
|
|
|
Thanks for your reply!!
Ok, in case if I set to use 50% of CPU cores.
Does it make sense to change to "6 cores per tasks" settings in PCs with 12/24 threads configuration or is better to stay with "12 cores per task"???
Is it normal that after changing to use 50% of CPU cores my Windows task manager shows only 60% of CPU usage not 100%??
____________
| |
|
|
|
Thanks for your reply!!
Ok, in case if I set to use 50% of CPU cores.
Does it make sense to change to "6 cores per tasks" settings in PCs with 12/24 threads configuration or is better to stay with "12 cores per task"???
Is it normal that after changing to use 50% of CPU cores my Windows task manager shows only 60% of CPU usage not 100%??
On your systems, the 2470 should be set to 8 threads/task, the 2698v3 also to 8 threads and the 5900X to 6. On multiprocessor (Xeons) and multi-die (Ryzen CCXes) cpus, settings should be calculated based on the individual chips. Going between CPUs or CCXes adds a ton of latency by using main memory which drastically slows down computation.
Your task manager reports are normal. It counts hyperthreading/SMT as extra cores and so when BOINC/PG is properly utilized at 50%, the total is always a bit over from various overheads and other processes.
____________
Eating more cheese on Thursdays. | |
|
|
|
|
ThrasherX-17 wrote: What is the priority.
The priority is to determine how many tasks a given computer can run at once without running into a memory access bottleneck. I don't know if there is a recent posting of the size of the "hot" data of llrCUL/llrWOO somewhere in this message board, but there is one over at anandtech.com. A rule of thumb is to divide a given CPU's last-level cache size by the FFT data size, and that's about how many LLR program instances should be started at once at most.
After this maximum number of concurrent tasks is known, the next and simpler problem is to choose the threadcount per task. The goal here is to fully utilize all physical cores. (To reach this goal, using Hyperthreads may or may not be beneficial. Though keep in mind that the higher the thread count per task, the higher the program-internal overhead for synchronization etc.)
Finally, you may want to deal with various oddities of some specific platforms: If you have Windows, you may run into trouble with enabled HT or SMT, let alone NUMA. Either find out workarounds for those, or simply use Linux. Or you may have a Zen...Zen 4 based AMD CPU with more than one core complex (CCX); in that case you might want to research how to deal with their cache segmentation (affects Windows and Linux). Or newest Intel small-socket CPUs with a mix of P and E cores might require further workarounds. | |
|
|
|
ThrasherX-17 wrote: What is the priority.
The priority is to determine how many tasks a given computer can run at once without running into a memory access bottleneck. I don't know if there is a recent posting of the size of the "hot" data of llrCUL/llrWOO somewhere in this message board, but there is one over at anandtech.com. A rule of thumb is to divide a given CPU's last-level cache size by the FFT data size, and that's about how many LLR program instances should be started at once at most.
After this maximum number of concurrent tasks is known, the next and simpler problem is to choose the threadcount per task. The goal here is to fully utilize all physical cores. (To reach this goal, using Hyperthreads may or may not be beneficial. Though keep in mind that the higher the thread count per task, the higher the program-internal overhead for synchronization etc.)
Finally, you may want to deal with various oddities of some specific platforms: If you have Windows, you may run into trouble with enabled HT or SMT, let alone NUMA. Either find out workarounds for those, or simply use Linux. Or you may have a Zen...Zen 4 based AMD CPU with more than one core complex (CCX); in that case you might want to research how to deal with their cache segmentation (affects Windows and Linux). Or newest Intel small-socket CPUs with a mix of P and E cores might require further workarounds.
I did a WOO it says FFT length is 2304 in the Stderr output file.
From what I understand you multiply the FFT size by 8. (No matter how many cores) That will give you the size of level cache 3 you need (Intel anyway, AMD has shared l3???)
So, 2304 x 8= 18,432 you want around 18mb of L3 cache.
That formula seems to work for me, my i9-9900k 8 core with 16mb cache seems slow
Going to fire up an E5-1660V3 with 8 cores 20mb L3 cache
I bet the older 8 core Xeon is faster than the newer 8 core 9900k with smaller cache. | |
|
|
|
|
I fired up an I7-5960x which is the same as an E5-1660V3 Xeon but for the two QPI ports for 2P. and ECC memory. It is an 8 core 20mb L3 @ 4ghz. Haswell E/EP
The I9-9900k is an 8 core 16mb L3 @ 4.4ghz Coffeelake.
We will see which is faster. Logic says the newer faster 9900k should be. So this should be a good test if the larger L3 makes the difference since I figure it needs 18mb of L3
If it's ok with PG I would like to see some of your combo and times, include the FFT size in the
Stderr output text, CPU model, cache and if you know it it's name IE Haswell, Coffeelake ect.
For WOO and CUL if you have it.
PS: Good luck everyone finding a prime. Hope someone does, it will be a big one for sure.
| |
|
|
|
|
bill1024 wrote: From what I understand you multiply the FFT size by 8.
Yep, that's because 8 Bytes (64 bits) wide numbers are being used in this transform.
bill1024 wrote: That will give you the size of level cache 3 you need (Intel anyway, AMD has shared l3???)
Intels have a unified level 3 cache, many with an "inclusive" L3$ policy, Skylake-X/-SP and newer big-socket Intels with "non-inclusive" L3$ policy. ("Inclusive" = always holds copies of L2$ contents. Obviously the different cache policies, maybe also different prefetcher algorithms and whatnot, affect how well a given cache size works for a given workload.) AMD Zen and later have non-inclusive level 3 caches, one L3$ per core complex (CCX). From what I understood, communication between CCXs goes through main memory.
In my experience, the performance of LLR2 goes gradually down when there is not enough processor cache, i.e. it does not drop drastically. (But that experience is with server processors with very large caches, on which one often still has a choice between slight under-use or slight over-use of the caches.)
Edit:
All this background aside, here is what I generally do before a PrimeGrid challenge:
I run a few short benchmarks which directly measure LLR2 performance with a given workunit. Those benchmarks are fully reproducible and enable me to find the optimum or near optimum config (for throughput optimum, or power efficiency optimum, or whatever) in a short timeframe and with high confidence. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Day 1 challenge stats (per subproject):
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-02 20:08:20 UTC)
5467 tasks have been sent out. [CPU/GPU/anonymous_platform: 5467 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1054 (19%) were aborted. [1054 (19%) / 0 (0%) / 0 (0%)]
63 (1%) came back with some kind of an error. [63 (1%) / 0 (0%) / 0 (0%)]
799 (15%) have returned a successful result. [799 (15%) / 0 (0%) / 0 (0%)]
3551 (65%) are still in progress. [3551 (65%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
375 (47%) are pending validation. [375 (47%) / 0 (0%) / 0 (0%)]
424 (53%) have been successfully validated. [424 (53%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22853898. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 0.63% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-02 20:08:20 UTC)
5891 tasks have been sent out. [CPU/GPU/anonymous_platform: 5891 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
365 (6%) were aborted. [365 (6%) / 0 (0%) / 0 (0%)]
101 (2%) came back with some kind of an error. [101 (2%) / 0 (0%) / 0 (0%)]
1768 (30%) have returned a successful result. [1768 (30%) / 0 (0%) / 0 (0%)]
3657 (62%) are still in progress. [3657 (62%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
715 (40%) are pending validation. [715 (40%) / 0 (0%) / 0 (0%)]
1053 (60%) have been successfully validated. [1053 (60%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=21990170. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 0.00% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
I run a few short benchmarks which directly measure LLR2 performance with a given workunit. Those benchmarks are fully reproducible and enable me to find the optimum or near optimum config (for throughput optimum, or power efficiency optimum, or whatever) in a short timeframe and with high confidence.
How do you run the benchmark? Could you please explain?
| |
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference!
Wow AVX512 performance is awesome!! Seems like the 13.752MB L3 cache size in i9-7900X doesn't affect the performance much. As a comparison my i7-6900K ETA for cullen task (8 core) is expected to finish in about 10 hours or so. It has 20MB of L3 with AVX2, more than i9-7900X (13.75MB). So really the AVX512 instruction set is the dominant factor here. I wonder what's the speed difference if both has the same AVX instruction set and how the L3 cache really affects the performance and perhaps the number of memory channels. Both supports quad memory channel. I have four sticks. | |
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference!
Wow AVX512 performance is awesome!! Seems like the 13.752MB L3 cache size in i9-7900X doesn't affect the performance much. As a comparison my i7-6900K ETA for cullen task (8 core) is expected to finish in about 10 hours or so. It has 20MB of L3 with AVX2, more than i9-7900X (13.75MB). So really the AVX512 instruction set is the dominant factor here. I wonder what's the speed difference if both has the same AVX instruction set and how the L3 cache really affects the performance and perhaps the number of memory channels. Both supports quad memory channel. I have four sticks.
The 7900x skylake have a larger L2 per core and somehow it can be added to the L3 cache for a larger total useful cache than it would appear to be, so can't really call it just 13mb, add in the L2 also, there is 1mb per core 10mb + the 13 = 23mb useful Lxx cache??? I do believe that's it.
xii5ku touched on skylake CPUs a little bit in his post about skylake and zen GPUs how they use cache.
I am pretty sure there are threads on here where it goes into better detail than I can explain.
But yeah in a nutshell AVX is awesome!!! It has to be the CPUs with 2 sets of AVX512 I believe some of the lesser CPUs have one instance of it.
EDIT: My 8 core 5960x haswell on a woo 2304FFT is at 68% in 6hrs with 2hr50min to go
11.1% per hour it says in the tasks properties.
just under 9 hrs vs your 10hr? But I not know what FFT size is your task.
And I have four 8gb sticks at 3600 17,21,21, something or other with the 7900x | |
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference!
Wow AVX512 performance is awesome!! Seems like the 13.752MB L3 cache size in i9-7900X doesn't affect the performance much. As a comparison my i7-6900K ETA for cullen task (8 core) is expected to finish in about 10 hours or so. It has 20MB of L3 with AVX2, more than i9-7900X (13.75MB). So really the AVX512 instruction set is the dominant factor here. I wonder what's the speed difference if both has the same AVX instruction set and how the L3 cache really affects the performance and perhaps the number of memory channels. Both supports quad memory channel. I have four sticks.
The 7900x skylake have a larger L2 per core and somehow it can be added to the L3 cache for a larger total useful cache than it would appear to be, so can't really call it just 13mb, add in the L2 also, there is 1mb per core 10mb + the 13 = 23mb useful Lxx cache??? I do believe that's it.
xii5ku touched on skylake CPUs a little bit in his post about skylake and zen GPUs how they use cache.
I am pretty sure there are threads on here where it goes into better detail than I can explain.
But yeah in a nutshell AVX is awesome!!! It has to be the CPUs with 2 sets of AVX512 I believe some of the lesser CPUs have one instance of it.
EDIT: My 8 core 5960x haswell on a woo 2304FFT is at 68% in 6hrs with 2hr50min to go
11.1% per hour it says in the tasks properties.
just under 9 hrs vs your 10hr? But I not know what FFT size is your task.
And I have four 8gb sticks at 3600 17,21,21, something or other with the 7900x
I see. My i7-6900K has 256KB per core, so L2+L3 = 22MB. My task should finish in about 2+ hours. Will report tonite on the FFT size. Looking at some of the closest task to mine probably 2400FFT which would still fit the L3 cache. BTW, what clock frequency are you running on the i9-7900K? Mine is set at 3.7GHz.
| |
|
|
|
|
My 7900x is running at 4ghz multiplier is 44-4AVX512 offset
The Haswel 5960x is at 4ghz also. No AVX offset. | |
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference!
The i7-5960x did the woo in 8.78 hours. Same size FFT
Somewhere around 1500 seconds faster. | |
|
|
|
My I9-9900k Coffeelake does the woo 33,000 9.1 hrs.
The 4.4 to 5 hour woos are on an I9-7900x 10 core with AVX512.
That makes a huge difference!
The i7-5960x did the woo in 8.78 hours. Same size FFT
Somewhere around 1500 seconds faster.
Cul FFT size is 2400K. It took 10.05 hours to finish. No AVX offset used for the 6900K.
| |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Day 2 stats!
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-03 09:02:03 UTC)
6672 tasks have been sent out. [CPU/GPU/anonymous_platform: 6672 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1066 (16%) were aborted. [1066 (16%) / 0 (0%) / 0 (0%)]
82 (1%) came back with some kind of an error. [82 (1%) / 0 (0%) / 0 (0%)]
1205 (18%) have returned a successful result. [1205 (18%) / 0 (0%) / 0 (0%)]
4319 (65%) are still in progress. [4319 (65%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
475 (39%) are pending validation. [475 (39%) / 0 (0%) / 0 (0%)]
730 (61%) have been successfully validated. [730 (61%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22907952. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 0.86% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-03 09:02:05 UTC)
7106 tasks have been sent out. [CPU/GPU/anonymous_platform: 7106 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
542 (8%) were aborted. [542 (8%) / 0 (0%) / 0 (0%)]
129 (2%) came back with some kind of an error. [129 (2%) / 0 (0%) / 0 (0%)]
2627 (37%) have returned a successful result. [2627 (37%) / 0 (0%) / 0 (0%)]
3808 (54%) are still in progress. [3808 (54%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
781 (30%) are pending validation. [781 (30%) / 0 (0%) / 0 (0%)]
1846 (70%) have been successfully validated. [1846 (70%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=21995720. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 0.03% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
|
Mr. Cool wrote: xii5ku wrote: I run a few short benchmarks which directly measure LLR2 performance with a given workunit. Those benchmarks are fully reproducible and enable me to find the optimum or near optimum config (for throughput optimum, or power efficiency optimum, or whatever) in a short timeframe and with high confidence.
How do you run the benchmark? Could you please explain?
What we normally do here at PG looks like this:
boinc
\_ primegrid wrapper
| \_ sllr2 -q21962175*2^21962175-1
\_ primegrid wrapper
| \_ sllr2 -q21979010*2^21979010-1
\_ primegrid wrapper
| \_ sllr2 -q22878096*2^22878096+1
\_ primegrid wrapper
\_ sllr2 -q21980895*2^21980895-1
My benchmarking setup simply looks like this:
shell script
\_ sllr2 -q22590179*2^22590179+1
\_ sllr2 -q22590179*2^22590179+1
\_ sllr2 -q22590179*2^22590179+1
\_ sllr2 -q22590179*2^22590179+1
The script can let the LLR workers run through the entire workunit, or stop them once a completion percentage of my choice is reached. Often enough, it is sufficient to stop a test after just a few percent. Since I test one and the same workunit over and over, I can make apples-to-apples comparisons between different choices of # of concurrent tasks, # of threads/task, various BIOS settings, and so on. I.e. the factor of variable workunit sizes is eliminated from these tests. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Day 3 stats:
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-04 07:47:08 UTC)
8378 tasks have been sent out. [CPU/GPU/anonymous_platform: 8378 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1137 (14%) were aborted. [1137 (14%) / 0 (0%) / 0 (0%)]
100 (1%) came back with some kind of an error. [100 (1%) / 0 (0%) / 0 (0%)]
2022 (24%) have returned a successful result. [2022 (24%) / 0 (0%) / 0 (0%)]
5119 (61%) are still in progress. [5119 (61%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
609 (30%) are pending validation. [609 (30%) / 0 (0%) / 0 (0%)]
1413 (70%) have been successfully validated. [1413 (70%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22983947. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 1.20% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-04 07:47:10 UTC)
9314 tasks have been sent out. [CPU/GPU/anonymous_platform: 9311 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
599 (6%) were aborted. [598 (6%) / 0 (0%) / 1 (0%)]
153 (2%) came back with some kind of an error. [153 (2%) / 0 (0%) / 0 (0%)]
4300 (46%) have returned a successful result. [4298 (46%) / 0 (0%) / 2 (0%)]
4263 (46%) are still in progress. [4263 (46%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1046 (24%) are pending validation. [1045 (24%) / 0 (0%) / 1 (0%)]
3254 (76%) have been successfully validated. [3253 (76%) / 0 (0%) / 1 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22067964. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 0.35% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Day 4 stats, separated by subproject:
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-05 17:56:08 UTC)
11812 tasks have been sent out. [CPU/GPU/anonymous_platform: 11812 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1386 (12%) were aborted. [1386 (12%) / 0 (0%) / 0 (0%)]
121 (1%) came back with some kind of an error. [121 (1%) / 0 (0%) / 0 (0%)]
3458 (29%) have returned a successful result. [3458 (29%) / 0 (0%) / 0 (0%)]
6847 (58%) are still in progress. [6847 (58%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
798 (23%) are pending validation. [798 (23%) / 0 (0%) / 0 (0%)]
2660 (77%) have been successfully validated. [2660 (77%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23144405. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 1.91% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-05 17:56:09 UTC)
13902 tasks have been sent out. [CPU/GPU/anonymous_platform: 13899 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
851 (6%) were aborted. [850 (6%) / 0 (0%) / 1 (0%)]
239 (2%) came back with some kind of an error. [239 (2%) / 0 (0%) / 0 (0%)]
6829 (49%) have returned a successful result. [6827 (49%) / 0 (0%) / 2 (0%)]
5985 (43%) are still in progress. [5985 (43%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1626 (24%) are pending validation. [1626 (24%) / 0 (0%) / 0 (0%)]
5203 (76%) have been successfully validated. [5201 (76%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22216404. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 1.03% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-07 05:18:00 UTC)
14683 tasks have been sent out. [CPU/GPU/anonymous_platform: 14683 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1565 (11%) were aborted. [1565 (11%) / 0 (0%) / 0 (0%)]
143 (1%) came back with some kind of an error. [143 (1%) / 0 (0%) / 0 (0%)]
4908 (33%) have returned a successful result. [4908 (33%) / 0 (0%) / 0 (0%)]
8067 (55%) are still in progress. [8067 (55%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
919 (19%) are pending validation. [919 (19%) / 0 (0%) / 0 (0%)]
3989 (81%) have been successfully validated. [3989 (81%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23270260. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 2.46% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-07 05:18:03 UTC)
17826 tasks have been sent out. [CPU/GPU/anonymous_platform: 17823 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
1203 (7%) were aborted. [1202 (7%) / 0 (0%) / 1 (0%)]
259 (1%) came back with some kind of an error. [259 (1%) / 0 (0%) / 0 (0%)]
9447 (53%) have returned a successful result. [9445 (53%) / 0 (0%) / 2 (0%)]
6917 (39%) are still in progress. [6917 (39%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1765 (19%) are pending validation. [1765 (19%) / 0 (0%) / 0 (0%)]
7682 (81%) have been successfully validated. [7680 (81%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22341930. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 1.60% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-08 06:07:12 UTC)
16698 tasks have been sent out. [CPU/GPU/anonymous_platform: 16698 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
1733 (10%) were aborted. [1733 (10%) / 0 (0%) / 0 (0%)]
158 (1%) came back with some kind of an error. [158 (1%) / 0 (0%) / 0 (0%)]
6110 (37%) have returned a successful result. [6110 (37%) / 0 (0%) / 0 (0%)]
8697 (52%) are still in progress. [8697 (52%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
960 (16%) are pending validation. [960 (16%) / 0 (0%) / 0 (0%)]
5150 (84%) have been successfully validated. [5150 (84%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23362299. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 2.86% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-08 06:07:15 UTC)
20332 tasks have been sent out. [CPU/GPU/anonymous_platform: 20329 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
1400 (7%) were aborted. [1399 (7%) / 0 (0%) / 1 (0%)]
272 (1%) came back with some kind of an error. [272 (1%) / 0 (0%) / 0 (0%)]
11334 (56%) have returned a successful result. [11332 (56%) / 0 (0%) / 2 (0%)]
7326 (36%) are still in progress. [7326 (36%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1544 (14%) are pending validation. [1544 (14%) / 0 (0%) / 0 (0%)]
9790 (86%) have been successfully validated. [9788 (86%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22420431. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 1.96% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
|
Thanks to everybody with the answers on my questions above.
Some things are too complicated and above my knowledge because I'm a doctor and anesthesiology is my field.
I sorted all of my PCs into 9 groups according to the number of cores and also GPU tasks and now when I set in boinc to use 50% of cores the computation times are a lot better.
The only problem that left is my Ryzen 9 3900X.
I bought it yesterday and this PC joined the challenge too.
I used the same settings as on 5900X build (12 cores per task) with use of 50% CPU cores but the computation times are awful.
Cullen task on 5900X - 6 hours 30 min
On 3900X - 14 hours and still 75% the cores temps are about 62-63 but when I start any different benchmark (Cinebench, passmark) the temps are near 85-86 degrees.
Memory is DDR4 2x16Gb @3200, XMP activated.
I tried to set 6 cores per task and the situation was even worse. 1,1% - 1 hour
Any ideas?
A little update:
I tried 2 workunits with (6 core per task) instead of 1 WU with 12 cores settings but when I suspended 1 task my computation times went from 1,1% - 1 hour to 5,8% - 1 hour.
____________
| |
|
|
|
Thanks to everybody with the answers on my questions above.
Some things are too complicated and above my knowledge because I'm a doctor and anesthesiology is my field.
I sorted all of my PCs into 9 groups according to the number of cores and also GPU tasks and now when I set in boinc to use 50% of cores the computation times are a lot better.
The only problem that left is my Ryzen 9 3900X.
I bought it yesterday and this PC joined the challenge too.
I used the same settings as on 5900X build (12 cores per task) with use of 50% CPU cores but the computation times are awful.
Cullen task on 5900X - 6 hours 30 min
On 3900X - 14 hours and still 75% the cores temps are about 62-63 but when I start any different benchmark (Cinebench, passmark) the temps are near 85-86 degrees.
Memory is DDR4 2x16Gb @3200, XMP activated.
I tried to set 6 cores per task and the situation was even worse. 1,1% - 1 hour
Any ideas?
A little update:
I tried 2 workunits with (6 core per task) instead of 1 WU with 12 cores settings but when I suspended 1 task my computation times went from 1,1% - 1 hour to 5,8% - 1 hour.
Couple thoughts/suggestions: If you're running Windows, do you have all the drivers installed and the BIOS updated? The AMD drivers/firmware help assign multithreaded tasks to single CCXes.
Next, what else is running at the same time? That is, what does your task manager say is eating resources, or what are the percentages that come up for the PG program? Is the "suspend computation when computer is in use" option accidentally checked?
____________
Eating more cheese on Thursdays. | |
|
|
|
[...]
The only problem that left is my Ryzen 9 3900X.
[...]
I used the same settings as on 5900X build (12 cores per task) with use of 50% CPU cores but the computation times are awful.
[...]
I tried 2 workunits with (6 core per task) instead of 1 WU with 12 cores settings but when I suspended 1 task my computation times went from 1,1% - 1 hour to 5,8% - 1 hour.
The Ryzen 5900x has a tiny little more power than the 3900x. ;-)
Also, both have their CPUs' level 3 caches divided up into two parts.
Any workunit that is not running solely in one part has efficiency constraints.
On Windows you can use Affinity Watcher by Pavel Atnashev to set the work straight into using only one part of the cache, on Linux you have to script it around taskset or run the work inside virtual machines that you put into only one part of the cache with taskset.
Iirc you could also use process lasso on Windows.
I'm on Linux using taskset on my virtual machines and am sufficiently happy with that.
____________
Greetings, Jens
147433824^131072+1 | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
One day left! Here are the stats.
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-09 08:55:17 UTC)
19023 tasks have been sent out. [CPU/GPU/anonymous_platform: 19023 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
2297 (12%) were aborted. [2297 (12%) / 0 (0%) / 0 (0%)]
169 (1%) came back with some kind of an error. [169 (1%) / 0 (0%) / 0 (0%)]
7493 (39%) have returned a successful result. [7493 (39%) / 0 (0%) / 0 (0%)]
9064 (48%) are still in progress. [9064 (48%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1150 (15%) are pending validation. [1150 (15%) / 0 (0%) / 0 (0%)]
6343 (85%) have been successfully validated. [6343 (85%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23449188. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 3.25% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-09 08:55:20 UTC)
24013 tasks have been sent out. [CPU/GPU/anonymous_platform: 24010 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
1932 (8%) were aborted. [1931 (8%) / 0 (0%) / 1 (0%)]
288 (1%) came back with some kind of an error. [288 (1%) / 0 (0%) / 0 (0%)]
13434 (56%) have returned a successful result. [13432 (56%) / 0 (0%) / 2 (0%)]
8359 (35%) are still in progress. [8359 (35%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1678 (12%) are pending validation. [1678 (12%) / 0 (0%) / 0 (0%)]
11756 (88%) have been successfully validated. [11754 (87%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22538066. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 2.49% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
|
gemini8 wrote: The Ryzen 5900x has a tiny little more power than the 3900x. ;-)
Also, both have their CPUs' level 3 caches divided up into two parts.
5900X: 2 last-level caches, 32 MB each
3900X: 4 last-level caches, 16 MB each
gemini8 wrote: or run the work inside virtual machines
There is no need to go that far. It can be done with either a single boinc client, or with 2 or more boinc clients all within the same host OS instance.
| |
|
|
|
|
Big thanks for all of your answers!
Computing times improved after updating BIOS (from 5% per 1 hour to 10% per 1 hour) 12 cores per WU and "use 50% of cores".
It seems to me it's not optimal performance of 3900X but a lot better that earlier.
10 hours for 1 Cullen LLR task
Also I've installed chipset drivers and Ryzen Master but didn't change anything there.
My running apps: boinc manager and teamviewer (0-0,5% of CPU usage)
____________
| |
|
|
|
|
Woohoo!!! One more day to go!!! | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-10 06:58:53 UTC)
20426 tasks have been sent out. [CPU/GPU/anonymous_platform: 20426 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
3201 (16%) were aborted. [3201 (16%) / 0 (0%) / 0 (0%)]
191 (1%) came back with some kind of an error. [191 (1%) / 0 (0%) / 0 (0%)]
8462 (41%) have returned a successful result. [8462 (41%) / 0 (0%) / 0 (0%)]
8572 (42%) are still in progress. [8572 (42%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1094 (13%) are pending validation. [1094 (13%) / 0 (0%) / 0 (0%)]
7368 (87%) have been successfully validated. [7368 (87%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23482338. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 3.39% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-10 06:58:55 UTC)
26663 tasks have been sent out. [CPU/GPU/anonymous_platform: 26660 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
2947 (11%) were aborted. [2946 (11%) / 0 (0%) / 1 (0%)]
303 (1%) came back with some kind of an error. [303 (1%) / 0 (0%) / 0 (0%)]
15093 (57%) have returned a successful result. [15091 (57%) / 0 (0%) / 2 (0%)]
8320 (31%) are still in progress. [8320 (31%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1779 (12%) are pending validation. [1779 (12%) / 0 (0%) / 0 (0%)]
13314 (88%) have been successfully validated. [13312 (88%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22595011. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 2.75% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
|
We want a prime! /JeppeSN | |
|
|
|
|
Yes!!!
Newest-Oldest Last Primes Found for:
Cullen: 2009-07-25 01:11:48 UTC
Woodall: 2018-03-21 22:13:39 UTC
BTW: the next newest-oldest prime found, after Cullen, in all of the categories listed in 'PrimeGrid Primes by Project' is:
Factorial: 2013-08-30 19:55:34 UTC | |
|
|
|
We want a prime! /JeppeSN
Seconded!
____________
| |
|
tngSend message
Joined: 29 Aug 10
Posts: 486
ID: 66603
Credit: 47,466,489,893
RAC: 27,436,496
                                                  
|
We want a prime! /JeppeSN
Seconded!
All in favor?
____________
 | |
|
|
|
|
Completed!!! | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
And we're done! Here are the final stats. Cleanup tasks are available now.
Challenge: Prime Meridian Day
App: 4 (CUL-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-11 07:07:32 UTC)
23313 tasks have been sent out. [CPU/GPU/anonymous_platform: 23313 (100%) / 0 (0%) / 0 (0%)]
Of those tasks that have been sent out:
7086 (30%) were aborted. [7086 (30%) / 0 (0%) / 0 (0%)]
202 (1%) came back with some kind of an error. [202 (1%) / 0 (0%) / 0 (0%)]
9417 (40%) have returned a successful result. [9417 (40%) / 0 (0%) / 0 (0%)]
6498 (28%) are still in progress. [6498 (28%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1269 (13%) are pending validation. [1269 (13%) / 0 (0%) / 0 (0%)]
8148 (87%) have been successfully validated. [8148 (87%) / 0 (0%) / 0 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=23491439. The leading edge was at n=22711734 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 3.43% as much as it had prior to the challenge!
Challenge: Prime Meridian Day
App: 3 (WOO-LLR)
Fast DC tasks are NOT included.
(As of 2022-11-11 07:07:36 UTC)
29671 tasks have been sent out. [CPU/GPU/anonymous_platform: 29668 (100%) / 0 (0%) / 3 (0%)]
Of those tasks that have been sent out:
4902 (17%) were aborted. [4901 (17%) / 0 (0%) / 1 (0%)]
304 (1%) came back with some kind of an error. [304 (1%) / 0 (0%) / 0 (0%)]
17034 (57%) have returned a successful result. [17032 (57%) / 0 (0%) / 2 (0%)]
7057 (24%) are still in progress. [7057 (24%) / 0 (0%) / 0 (0%)]
Of the tasks that have been returned successfully:
1935 (11%) are pending validation. [1935 (11%) / 0 (0%) / 0 (0%)]
15099 (89%) have been successfully validated. [15097 (89%) / 0 (0%) / 2 (0%)]
0 (0%) were invalid. [0 (0%) / 0 (0%) / 0 (0%)]
0 (0%) are inconclusive. [0 (0%) / 0 (0%) / 0 (0%)]
The current leading edge (i.e., latest work unit for which work has actually been sent out to a host) is n=22642964. The leading edge was at n=21990170 at the beginning of the challenge. Since the challenge started, the leading edge has advanced 2.97% as much as it had prior to the challenge!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
|
|
|
Clean up tasks? | |
|
Dave
Send message
Joined: 13 Feb 12
Posts: 3209
ID: 130544
Credit: 2,288,005,894
RAC: 763,651
|
Clean up tasks?
Clean-up is automatic via normal daily production. | |
|
|
|
[...]
BTW: the next newest-oldest prime found, after Cullen, in all of the categories listed in 'PrimeGrid Primes by Project' is:
Factorial: 2013-08-30 19:55:34 UTC
However, for factorial (FPS) a lot of instances have been found outside PrimeGrid; see Top 20 Factorial Primes (or this search). /JeppeSN | |
|
|
|
|
Thank you to everyone at the backend of this challenge, for all of your efforts!
And to all that participated.
This was my 1st challenge and was a fun break from the normal searches! I'm really looking forward to upcoming challenges :)
Cheers everyone!!! | |
|
|
|
|
Thanks! | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
Cleanup Status:
Nov 12: 1176 tasks outstanding; 675 affecting individual (159) scoring positions; 62 affecting team (17) scoring positions.
Nov 18: 90 tasks outstanding; 34 affecting individual (25) scoring positions; 4 affecting team (3) scoring positions.
Nov 19: 60 tasks outstanding; 25 affecting individual (18) scoring positions; 3 affecting team (2) scoring positions.
Nov 21: 39 tasks outstanding; 19 affecting individual (12) scoring positions; 2 affecting team (1) scoring positions.
Nov 22: 21 tasks outstanding; 9 affecting individual (5) scoring positions; 2 affecting team (1) scoring positions.
Nov 23: 9 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 24: 8 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 25: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 26: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 27: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 28: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 29: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Nov 30: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Dec 01: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Dec 02: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Dec 03: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
Dec 04: 7 tasks outstanding; 4 affecting individual (1) scoring positions; 0 affecting team (0) scoring positions.
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
Michael GutierrezVolunteer moderator
Project administrator
Project scientist
Send message
Joined: 21 Mar 17
Posts: 349
ID: 764476
Credit: 46,281,510
RAC: 0
|
|
The results are final!
During the 10 day challenge, we completed more than 50,000 tasks (including doublechecks).
127 teams and 577 individuals participated in the challenge.
Top Three Individuals:
1. vaughan with 2,921 tasks completed
2. tng with 1,401 tasks completed
3. markfw with 1,958 tasks completed
Top Three Teams:
1. TeAm AnandTech with 10,397 tasks completed
2. Czech National Team with 5,153 tasks completed
3. Antarctic Crunchers with 6,725 tasks completed
The final challenge of the year starts in just a few days! Happy crunching!
____________
Time flies like an arrow. Fruit flies like a banana. | |
|
Post to thread
Message boards :
Number crunching :
Prime Meridian Day Challenge |
