Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Generalized Fermat Prime Search :
GFN-15 software upgraded to Genefer22
Author |
Message |
Michael Goetz Volunteer moderator Project administrator
 Send message
Joined: 21 Jan 10 Posts: 14037 ID: 53948 Credit: 476,993,571 RAC: 281,271
                               
|
We have updated the GFN-15 apps to the Genefer22, the big 2022 upgrade. Note that we are NOT yet doing fast DC on GFN-15, but running Genefer22 does provide these advantages:
A higher B limit for CPU tasks, which are once again available.
Support for multi-threading on CPU apps, although it's not likely to be very useful on such small tasks.
Support for discreet Intel ARC GPUs.
Support for Mac M1 CPUs.
Support for Mac M1 GPUs.
____________
My lucky number is 75898524288+1 | |
|
Monkeydee Volunteer tester
 Send message
Joined: 8 Dec 13 Posts: 548 ID: 284516 Credit: 1,706,339,811 RAC: 3,240,363
                            
|
There seems to be a bit of an issue.
Nvidia cards don't seem to be validating with other nvidia cards.
Examples:
https://www.primegrid.com/workunit.php?wuid=907049153
https://www.primegrid.com/workunit.php?wuid=907036064
https://www.primegrid.com/workunit.php?wuid=906980340
The workunits aren't validating until after a CPU or AMD GPU task is returned and then we end up with many valid tasks per workunit
____________
My Primes
Badge Score: 4*2 + 6*2 + 7*1 + 8*11 + 9*1 + 11*3 + 12*1 = 169
| |
|
|
I know I'm not a big (nor steady) cruncher on PrimeGrid but I love the project and it happens I own a nvidia card so... I'll keep an eye on this :)
Thanks for the update! | |
|
Michael Goetz Volunteer moderator Project administrator
 Send message
Joined: 21 Jan 10 Posts: 14037 ID: 53948 Credit: 476,993,571 RAC: 281,271
                               
|
There seems to be a bit of an issue.
Nvidia cards don't seem to be validating with other nvidia cards.
Examples:
https://www.primegrid.com/workunit.php?wuid=907049153
https://www.primegrid.com/workunit.php?wuid=907036064
https://www.primegrid.com/workunit.php?wuid=906980340
The workunits aren't validating until after a CPU or AMD GPU task is returned and then we end up with many valid tasks per workunit
Noted.
____________
My lucky number is 75898524288+1 | |
|
Michael Goetz Volunteer moderator Project administrator
 Send message
Joined: 21 Jan 10 Posts: 14037 ID: 53948 Credit: 476,993,571 RAC: 281,271
                               
|
For the time being, I've turned off the new Nvidia apps. This fixes the problem of too many tasks being required for validations. All the other new apps are still enabled, so support for CPUs, Intel ARC, M1, and ARM remain enabled.
This does mean, however, that the original bug this was intended to fix is back. But if you weren't affected by it before, you won't be affected by it now.
____________
My lucky number is 75898524288+1 | |
|
|
I have 3 PCs, all running Windows 10 Professional. Since GFN-15 was upgraded to Genefer22, on one of my 3 PCs, all the CPU GFN tasks (GFN-15 and GFN-16) are running very slowly. For example, one GFN-15 CPU task, running with 3 CPUs, ran for 25,979 seconds (7+ hours), the CPU Time was 198 seconds and the task was only 70% complete. All other non-GFN tasks are running normally on this PC. I've rebooted the PC multiple times, and there has been no improvement with CPU GFN run times. (CPU GFN-15 and GFN-16 tasks are running fine on my other 2 PCs).
What action do you recommend I take to try to fix this problem? | |
|
Yves Gallot Volunteer developer Project scientist Send message
Joined: 19 Aug 12 Posts: 843 ID: 164101 Credit: 306,521,622 RAC: 5,385

|
I have 3 PCs, all running Windows 10 Professional. Since GFN-15 was upgraded to Genefer22, on one of my 3 PCs, all the CPU GFN tasks (GFN-15 and GFN-16) are running very slowly. For example, one GFN-15 CPU task, running with 3 CPUs, ran for 25,979 seconds (7+ hours), the CPU Time was 198 seconds and the task was only 70% complete. All other non-GFN tasks are running normally on this PC. I've rebooted the PC multiple times, and there has been no improvement with CPU GFN run times. (CPU GFN-15 and GFN-16 tasks are running fine on my other 2 PCs).
What action do you recommend I take to try to fix this problem?
i5-2520M, i5-3210M and i5-7200U have 4 logical cores but 2 physical cores (and 4 threads).
The tasks will run faster if a single task is tested and the max number of threads for each task is 2 (and not 3 or 4).
| |
|
|
Changing from 3 to 2 threads fixed all the problems. Thank you!!!! | |
|
|
Are those requirements still valid for GFN GPU tasks?
Because my GTX 275 gets this error (which is not send back to the server for some reason):
geneferg version 22.12.2 (win64, gcc-12.2.0, boinc-7.20.2)
Copyright (c) 2022, Yves Gallot
genefer is free source code, under the MIT license.
Command line: '-boinc -p -n 18 -b 24971434 -f gproof'
Running on device 'GeForce GTX 275', vendor 'NVIDIA Corporation', version 'OpenCL 1.0 CUDA', driver '342.01'ptxas error : Entry function 'fwd2048p' uses too much shared data (0x6020 bytes, 0x4000 max)
ptxas error : Entry function 'backward1024' uses too much shared data (0x6028 bytes, 0x4000 max)
ptxas error : Entry function 'mul2048' uses too much shared data (0x6028 bytes, 0x4000 max)
ptxas error : Entry function 'square2048' uses too much shared data (0x6020 bytes, 0x4000 max)
ptxas error : Entry function 'forward1024' uses too much shared data (0x6028 bytes, 0x4000 max)
Error: opencl error: CL_INVALID_BINARY.
Same error for GFN18 and GFN22.
____________
.
| |
|
Yves Gallot Volunteer developer Project scientist Send message
Joined: 19 Aug 12 Posts: 843 ID: 164101 Credit: 306,521,622 RAC: 5,385

|
This page needs to be updated.
The requirements are defined in Edit PrimeGrid preferences.
Notes for all Generalized Fermat applications:
Supported platforms:
- Windows: CPU1: x86, x64. GPU2: Nvidia, AMD, Intel ARC.
- Linux: CPU1: x86, x64, ARM64. GPU2: Nvidia, AMD, Intel ARC.
- Mac: CPU1: x64, ARM64. GPU2: Nvidia, AMD, Apple M-Series.
- For CPU tasks, multi-threading is supported. Click here to set the maximum number of threads.
- For GFN-21 CPU and GFN-22 CPU tasks, multi-threading is turned on automatically. A minimum of 4 CPU logical cores are required.
- Except for GFN-15, all GFN use fast proof tasks so no double check tasks are needed. Everyone is "first"!
1 CPU tasks are not available for "Do You Feel Lucky?".
2 OpenCL 1.1 or higher is required.
OpenCL 1.0 is not supported. | |
|
|
Haven't seen that, thanks, thought the server simply won't send me anything my GPU can't crunch.
____________
.
| |
|
Post to thread
Message boards :
Generalized Fermat Prime Search :
GFN-15 software upgraded to Genefer22 |