PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
1) Message boards : Number crunching : LLR initialization (Message 150316)
Posted 9 days ago by Profile composite
To give Yves' answer a concrete example,
the time saved by precomputing 21 iterations of PPS-MEGA
for Intel(R) Core(TM)2 Duo CPU E8400 @ 3.00GHz,
which takes about 11,200 CPU seconds using 2 threads to do the full main task,
would be around .06 CPU seconds.
2) Message boards : Number crunching : LLR initialization (Message 150314)
Posted 9 days ago by Profile composite
Thanks Yves, very enlightening.
Quoting one of my professors, back in the day...
That's shooting at mice, we're after elephants.
3) Message boards : Number crunching : LLR initialization (Message 150302)
Posted 10 days ago by Profile composite
How much compute time elapses from the start of the LLR algorthm until the computation becomes larger than the smallest untested candidate's modulus in each subproject? Before that point computations are the same for all workunits in the subproject, so it would save CPU time globally if the server issues workunits which are precomputed to start from the same advanced step of LLR instead of the first step.

EDIT: A precomputed advanced starting point would be particularly efficient for CPUs receiving workunits that don't have instructions like FMA.
4) Message boards : Number crunching : Tesla anyone? (Message 150138)
Posted 26 days ago by Profile composite
I've just written to my MP asking him to urge the Canadian Government to set energy efficiency standards for cryptocurrency, arguing that it will help achieve carbon footprint targets, as well as reduce the pressure on electricity prices. As the market price for crypto goes up, miners will pay anything for electricity and drive up the cost for the rest of us.

As a byproduct of energy efficiency standards for crypto, quantities of fast GPUs would be dumped on the secondary market with no competition from industrial-scale miners.

EDIT: An example of the shift of mining capacity from China to North America.
5) Message boards : Number crunching : Bang for the buck? (Message 150133)
Posted 26 days ago by Profile composite
Let's be more clear about this.

The question is always rent or buy if you're not using the surplus capacity of some other owner's GPU farm. Six years down the road your RTX 3090 is going to look as dated as the GTX 760 I'm currently running, and you'll have spent an average of $300 per year on the acquisition cost alone.

NVidia's newest CPU+GPU line "Grace" is data-center only. It won't be the end-of-the-road for consumer-grade GPUs for a while, just as chip makers still make enthusiast-level CPUs. Their price however is going through the roof as there will be stiff competition for dwindling supply. GPU volume is shifting to mobile and cloud as AI is ramping up to become the dominant application for GPUs. Sooner or later the rent-or-buy question will be about datacenter-based GPU, whether it's in the public cloud or some organization's compute resource.

The largest contributors to PG on GPU subprojects are either already renting cloud-based GPUs or they are relying on fleets of PCs set up for business or education. The fleets will evaporate in the next round of hardware refresh as computing moves to the cloud at scale simply because of datacenter economics.

So at some point PrimeGrid members who want to own GPU capacity will have to decide whether to spend ever more money on a top-of-the-line consumer-grade GPU, or a buy a share of a datacenter-based resource. The question hasn't been broached before, but this is certain to happen within the expected lifetime of many GPU subprojects.
6) Message boards : Number crunching : Bang for the buck? (Message 150120)
Posted 27 days ago by Profile composite
Benchmarks are required to make an assessment.

A100 performs similar to RTX3090 when it's processing bandwidth-limited computations like FFT.

For example, this series of benchmarks doesn't show huge throughput improvement over RTX3090 when the A100 is sitting on a PCIe bus for some workloads.
7) Message boards : Number crunching : Bang for the buck? (Message 150117)
Posted 28 days ago by Profile composite
Would it make financial sense to crowdfund the purchase and operation of a datacenter-scale GPU dedicated to PrimeGrid? Not especially advocating this particular brand which has 4 NVidia A100 GPUs for US$65,000 (there are other brands), compared to buying and running a personal high end NVidia GPU and computer to feed it.

Would it be feasible to divide up the subproject credit (and prime reporting) to the funders as single computing resource, or use it as a virtual GPU for each of the funders?

Would there be resale value in the contributions, i.e. a secondary market for the machine fraction?
8) Message boards : Number crunching : Yuri's Night Challenge (Message 150034)
Posted 32 days ago by Profile composite
This is propably a stupid question but will there be special badges awarded for Yuri's Night Challange or will credits be counted towards WW?

Badges are not awarded for challenges. However all work you do earns credit for the subproject badges.
9) Message boards : Number crunching : WW & 3080 ranking? (Message 150029)
Posted 33 days ago by Profile composite
I noticed in the "computer types" link in the menu at bottom left, that about 2/3 of systems running PrimeGrid are Linux. They outnumber Windows systems by about 2:1 ratio.
10) Message boards : Number crunching : Too much work when joining (Message 149889)
Posted 38 days ago by Profile composite

3) It's unusual to go from single channel RAM to dual channel RAM by adding a single stick due to the difficulties in getting matched memory sticks. RAM is usually sold as a multi-channel package because the vendor ensures the RAM timing is the same across all sticks. This kind of upgrade might even cause your computer to go slower without fine-tuning. See the aformentioned article about this bugaboo.
Not sure why people always say that. You just buy any stick with the same CAS. Even if you don't, the BIOS/MB just runs them at the lesser of the two speeds. I've got some machines with totally mismatched RAM, they just run as though I had two of the slower ones. Same goes for striping hard disks in a RAID.

I am inclined to agree with you on theoretical grounds but in the general case the advice has to be conservative when not inserting caveats and disclaimers everywhere. That's a slightly aged article so the author is probably referring to an older BIOS behaviour - just ignoring the XMP settings altogether and dropping the RAM clock to the minimum JEDEC spec just to make sure the system is bootable. I have a 7 year old X99 mobo which does exactly that even with a proper matched set of RAM if it doesn't like what I've done with a custom RAM setting.


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2021 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 5.84, 6.32, 5.85
Generated 15 May 2021 | 12:01:42 UTC