7 ms·

You can of course just run folding@home on your CPU as well, no nvidia requirement... required.

by SCdF 7y ago

KarlKemp 7y ago
It’s been a while since I worked in protein folding and tried folding@home, but IIRC the difference between GPUs and CPUs was at about the scale we see in machine learning, something like 20x to 200x. That makes the usefulness of CPUs for this purpose questionable—you might be contributing more to that other disaster facing humanity than alleviating the current one.
- slizard 7y ago
  [Disclaimer: How do I know this stuff? I'm a core developer of GROMACS, one of the major molecular dynamics simulation FOSS community codes. Molecular dynamic is the algorithm/method behind the F@H simulations. I work as a researcher in high performance computing and have (co-)developed many of the parallel algorithms and GPU acceleration in GROMACS.] 20-200x is simply not true. Typically, such numbers are a result of comparing unoptimized CPU code to moderately or well-optimized GPU code which is often misleading. (Such differences are however perfectly reasonable when comparing hardware-accelerated workloads like ML/DL). If you compare actually well-optimized codes, you'll see more like ~4-5x difference in performance for FLOP/instrction-bound code as it is the case for well-optimized molecular dynamics. Case in point, I recently pointed out the huge difference in CPU performance of two of the top molecular simulation codes, one of which is 8-10x faster on CPUs than the other, solving the same problem [2]. F@H relies on GROMACS as a CPU engine [1] which happens to be the same code as I quoted above as the fast one. The trouble is that F@H has not updated their CPU engine for many years and distribute CPU binaries which lack crucial SIMD optimizations to allow making use of AVX2/AVX512 on modern x86 CPUs as well as the years of algorithmic improvement and code optimization we made. These two factors combined lead to _significantly_ lower F@H CPU performance compared to what they had we're they using a recent GROMACS engine. Consequently, due to the combination of an inherent performance advantage of GPUs and the severely outdated CPU engine, it is indeed not worth wasting energy with running F@H on CPUs. [1] https://en.wikipedia.org/wiki/List_of_Folding@home_cores#GROMACS https://en.wikipedia.org/wiki/List_of_Folding@home_cores#GRO... [2] https://twitter.com/twilard/status/1235142089156984832?s=20 https://twitter.com/twilard/status/1235142089156984832?s=20 Edit 1: adjusted wording to reflect that the performance difference between running outdated GROMACS version and subotimal SIMD optimizations on modern hw can have a range of performance difference, depending on hardware and inputs. Edit 2: fixed typo + formatting.
  - swinglock 7y ago
    It seems irresponsible to waste donated cycles like that. Does anyone else do it better?
    - slizard 7y ago
      Can you please elaborate? Are you referring to the F@H CPU "cores" being outdated and inefficient?
      - swinglock 7y ago
        Yes, if it's so. When donating money or compute time, I want my donation to be used efficiently and not be wasted by high overheads.
        slizard 7y ago
        > When donating money or compute time, I want my donation to be used efficiently and not be wasted by high overheads. Fair point. I think this is something you should bring up with the authors of Folding@Home. I do not work on that project. My personal view on this is that there is always a cost/benefit balance that one has to strike which is often tricky especially given the considerably constrained resources as it is typically the case in academic computational/simulation tool development. It is however, as you point out, a great responsibility of the researchers and developers of codes to make sure that choices made and action taken (or not) do not lead to disproportionate waste of resources donated by volunteers or awarded through a grant by research funding agencies. I do not know the detailed reasons why F@H chose to not update CPU "core" since FahCore_a7 (AFAIK based one GROMACS code from 2014), but it is likely related to the aforementioned cost/benefit analysis done in their team. One of the motivations could have been (just hypothesizing) that the software engineering efforts estimated to be required to update FahCore for CPUs (which generate a very small fraction of the "points") would have taken away resources from the GPU FahCore.
    - slizard 7y ago
      > Does anyone else do it better? To that question, assuming by "anyone" here you are asking about other donate-your-cycles-distributed-computing-projects: I am not too familiar with how well-optimized the codes of different @home projects are. Taking a few steps back, perhaps the efficiency of these codes is the lesser issue and to be honest, in some (many?) cases other forms of donation/contribution may further more scientific progress than simply crunching numbers on one's home PC.
llukas 7y ago
https://foldingathome.org/faqs/high-performance/folding-graphics-cards/gpus-can-help-fah/ https://foldingathome.org/faqs/high-performance/folding-grap... GPUs are faster.
micael_dias 7y ago
From the reddit thread: "For people worried about electricity usage: My 3900x/2070s system, if using both cpu+gpu for folding full, uses 430w at the wall plug. Even at 21cents per kw/h this is about 9 cents/h. When setting up your machine, it's best to go to slots and remove your cpu. GPU tends to make all the points, especially if you have a newer GPU. I removed my 12 core 3900x from slots and barely noticed a drop in points while my system was running much cooler and with even less power usage. Leave it to default disease and you'll likely get covid-19 projects. I use both my 2070s cards and I get estimated 3.2M ppd for $2/24hrs of electricity. TLDR: PC's don't use much power. Get folding@home and just use your GPU(s) at full. Leave it to default to get covid-19 projects."
- sundvor 7y ago
  Just as with crypto mining, I'd recommend to set up a new Folding profile on your GPU settings. You want to be controlling GPU power limits eg reduce it by a bit and adding a more aggressive fan setting. I applied this strategy when bitcoin mining on my 1080ti which I got shortly after release. I was aiming for sub 60c temps. The results paid for the GPU, and I didn't fry it either - still my daily driver today. (The 1080TI is / was such a beast, with exceptional value for money.)
m4rtink 7y ago
Filding@home is not Nvidia obly (unlike other GPGPU projects, unfortunately). Works just fine on my AMD card.
cbg0 7y ago
At this time they are not sending covid CPU projects to clients, only GPU ones.
- DelightOne 7y ago
  I only got a CPU project, but no GPU (GTX 1080). > No WUs available for this configuration
  - SpikedCola 7y ago
    Same problem on my GTX 1050 (just tried updating drivers, no change). However, my laptop's GTX460M has pulled down project 11745 and is folding away. EDIT after fiddling with gpu-index, cuda-index, and opencl-index (had to manually set them to point to my GTX 1050 [also have a BARTS card that is unsupported]) I was able to get a WU to download :)
    - nocture 7y ago
      how do you find the correct index values? My 2080 Ti receives no work.
      - SpikedCola 7y ago
        In my log window I see: GPUs: 2 GPU 0: Bus:1 Slot:0 Func:0 AMD:4 Barts XT [Radeon HD 6800 Series] GPU 1: Bus:2 Slot:0 Func:0 NVIDIA:7 GP107 [GeForce GTX 1050 LP] 1862 CUDA Device 0: Platform:0 Device:0 Bus:2 Slot:0 Compute:6.1 Driver:8.0 OpenCL Device 0: Platform:0 Device:0 Bus:2 Slot:0 Compute:1.2 Driver:378.49 OpenCL Device 1: Platform:1 Device:0 Bus:1 Slot:0 Compute:1.2 Driver:1800.8 I want to use the GTX1050, so I set gpu-index to 1. Only 1 CUDA device so I set cuda-index to 0. To use the OpenCL device for the GTX, I looked at the bus # (bus 2), found OpenCL device on bus 2 (device 0), so I set opengl-index to 0. Not sure if this is the right way, but it worked for me.
- Leo_Verto 7y ago
  That's not true anymore, there are CPU projects now: https://foldingforum.org/viewtopic.php?f=24&t=32339 https://foldingforum.org/viewtopic.php?f=24&t=32339 https://foldingforum.org/viewtopic.php?f=24&t=32341 https://foldingforum.org/viewtopic.php?f=24&t=32341

You can of course just run folding@home on your CPU as well, no nvidia requirement... required.

by SCdF 7y ago

You can of course just run folding@home on your CPU as well, no nvidia requirement... required.

It’s been a while since I worked in protein folding and tried folding@home, but IIRC the difference between GPUs and CPUs was at about the scale we see in machine learning, something like 20x to 200x. That makes the usefulness of CPUs for this purpose questionable—you might be contributing more to that other disaster facing humanity than alleviating the current one.

slizard 7y ago
[Disclaimer: How do I know this stuff? I'm a core developer of GROMACS, one of the major molecular dynamics simulation FOSS community codes. Molecular dynamic is the algorithm/method behind the F@H simulations. I work as a researcher in high performance computing and have (co-)developed many of the parallel algorithms and GPU acceleration in GROMACS.] 20-200x is simply not true. Typically, such numbers are a result of comparing unoptimized CPU code to moderately or well-optimized GPU code which is often misleading. (Such differences are however perfectly reasonable when comparing hardware-accelerated workloads like ML/DL). If you compare actually well-optimized codes, you'll see more like ~4-5x difference in performance for FLOP/instrction-bound code as it is the case for well-optimized molecular dynamics. Case in point, I recently pointed out the huge difference in CPU performance of two of the top molecular simulation codes, one of which is 8-10x faster on CPUs than the other, solving the same problem [2]. F@H relies on GROMACS as a CPU engine [1] which happens to be the same code as I quoted above as the fast one. The trouble is that F@H has not updated their CPU engine for many years and distribute CPU binaries which lack crucial SIMD optimizations to allow making use of AVX2/AVX512 on modern x86 CPUs as well as the years of algorithmic improvement and code optimization we made. These two factors combined lead to _significantly_ lower F@H CPU performance compared to what they had we're they using a recent GROMACS engine. Consequently, due to the combination of an inherent performance advantage of GPUs and the severely outdated CPU engine, it is indeed not worth wasting energy with running F@H on CPUs. [1] https://en.wikipedia.org/wiki/List_of_Folding@home_cores#GROMACS https://en.wikipedia.org/wiki/List_of_Folding@home_cores#GRO... [2] https://twitter.com/twilard/status/1235142089156984832?s=20 https://twitter.com/twilard/status/1235142089156984832?s=20 Edit 1: adjusted wording to reflect that the performance difference between running outdated GROMACS version and subotimal SIMD optimizations on modern hw can have a range of performance difference, depending on hardware and inputs. Edit 2: fixed typo + formatting.
- swinglock 7y ago
  It seems irresponsible to waste donated cycles like that. Does anyone else do it better?
  - slizard 7y ago
    Can you please elaborate? Are you referring to the F@H CPU "cores" being outdated and inefficient?
    - swinglock 7y ago
      Yes, if it's so. When donating money or compute time, I want my donation to be used efficiently and not be wasted by high overheads.
      - slizard 7y ago
        > When donating money or compute time, I want my donation to be used efficiently and not be wasted by high overheads. Fair point. I think this is something you should bring up with the authors of Folding@Home. I do not work on that project. My personal view on this is that there is always a cost/benefit balance that one has to strike which is often tricky especially given the considerably constrained resources as it is typically the case in academic computational/simulation tool development. It is however, as you point out, a great responsibility of the researchers and developers of codes to make sure that choices made and action taken (or not) do not lead to disproportionate waste of resources donated by volunteers or awarded through a grant by research funding agencies. I do not know the detailed reasons why F@H chose to not update CPU "core" since FahCore_a7 (AFAIK based one GROMACS code from 2014), but it is likely related to the aforementioned cost/benefit analysis done in their team. One of the motivations could have been (just hypothesizing) that the software engineering efforts estimated to be required to update FahCore for CPUs (which generate a very small fraction of the "points") would have taken away resources from the GPU FahCore.
  - slizard 7y ago
    > Does anyone else do it better? To that question, assuming by "anyone" here you are asking about other donate-your-cycles-distributed-computing-projects: I am not too familiar with how well-optimized the codes of different @home projects are. Taking a few steps back, perhaps the efficiency of these codes is the lesser issue and to be honest, in some (many?) cases other forms of donation/contribution may further more scientific progress than simply crunching numbers on one's home PC.

llukas 7y ago

https://foldingathome.org/faqs/high-performance/folding-graphics-cards/gpus-can-help-fah/ https://foldingathome.org/faqs/high-performance/folding-grap... GPUs are faster.

micael_dias 7y ago

From the reddit thread: "For people worried about electricity usage: My 3900x/2070s system, if using both cpu+gpu for folding full, uses 430w at the wall plug. Even at 21cents per kw/h this is about 9 cents/h. When setting up your machine, it's best to go to slots and remove your cpu. GPU tends to make all the points, especially if you have a newer GPU. I removed my 12 core 3900x from slots and barely noticed a drop in points while my system was running much cooler and with even less power usage. Leave it to default disease and you'll likely get covid-19 projects. I use both my 2070s cards and I get estimated 3.2M ppd for $2/24hrs of electricity. TLDR: PC's don't use much power. Get folding@home and just use your GPU(s) at full. Leave it to default to get covid-19 projects."

sundvor 7y ago
Just as with crypto mining, I'd recommend to set up a new Folding profile on your GPU settings. You want to be controlling GPU power limits eg reduce it by a bit and adding a more aggressive fan setting. I applied this strategy when bitcoin mining on my 1080ti which I got shortly after release. I was aiming for sub 60c temps. The results paid for the GPU, and I didn't fry it either - still my daily driver today. (The 1080TI is / was such a beast, with exceptional value for money.)

m4rtink 7y ago

Filding@home is not Nvidia obly (unlike other GPGPU projects, unfortunately). Works just fine on my AMD card.

cbg0 7y ago

At this time they are not sending covid CPU projects to clients, only GPU ones.

DelightOne 7y ago
I only got a CPU project, but no GPU (GTX 1080). > No WUs available for this configuration
- SpikedCola 7y ago
  Same problem on my GTX 1050 (just tried updating drivers, no change). However, my laptop's GTX460M has pulled down project 11745 and is folding away. EDIT after fiddling with gpu-index, cuda-index, and opencl-index (had to manually set them to point to my GTX 1050 [also have a BARTS card that is unsupported]) I was able to get a WU to download :)
  - nocture 7y ago
    how do you find the correct index values? My 2080 Ti receives no work.
    - SpikedCola 7y ago
      In my log window I see: GPUs: 2 GPU 0: Bus:1 Slot:0 Func:0 AMD:4 Barts XT [Radeon HD 6800 Series] GPU 1: Bus:2 Slot:0 Func:0 NVIDIA:7 GP107 [GeForce GTX 1050 LP] 1862 CUDA Device 0: Platform:0 Device:0 Bus:2 Slot:0 Compute:6.1 Driver:8.0 OpenCL Device 0: Platform:0 Device:0 Bus:2 Slot:0 Compute:1.2 Driver:378.49 OpenCL Device 1: Platform:1 Device:0 Bus:1 Slot:0 Compute:1.2 Driver:1800.8 I want to use the GTX1050, so I set gpu-index to 1. Only 1 CUDA device so I set cuda-index to 0. To use the OpenCL device for the GTX, I looked at the bus # (bus 2), found OpenCL device on bus 2 (device 0), so I set opengl-index to 0. Not sure if this is the right way, but it worked for me.
Leo_Verto 7y ago
That's not true anymore, there are CPU projects now: https://foldingforum.org/viewtopic.php?f=24&t=32339 https://foldingforum.org/viewtopic.php?f=24&t=32339 https://foldingforum.org/viewtopic.php?f=24&t=32341 https://foldingforum.org/viewtopic.php?f=24&t=32341