Folding@Home Points Per Day results

2

Comments

  • ardichokeardichoke Icrontian
    edited June 2011
    Dual Xeon E5645, 12 cores (HyperThreaded) @ 2400MHz, Debian, SMP2 - 42617 ppd
  • ardichokeardichoke Icrontian
    edited June 2011
    Dual Xeon E5645, 12 cores (HyperThreaded) @ 2400MHz, Debian, bigadv - 64318 ppd
  • _k_k P-Town, Texas Icrontian
    edited June 2011
    Are you simply benching or letting these things run a full WU?
  • ardichokeardichoke Icrontian
    edited June 2011
    I'm letting them run a single full WU. I created a Storm on Demand image that builds, boots, fires up folding with my info and the -oneunit flag, automatically adds -smp and -bigadv depending on the number of cores. I then go in and remove the server once it has finished it's WU and I have collected my data.
  • ardichokeardichoke Icrontian
    edited July 2011
    AMD Phenom II X6 1055T @ 2812MHz, Debian, SMP - 8322 ppd
  • ardichokeardichoke Icrontian
    edited July 2011
    Intel Dual Xeon E5506, 8 cores @ 2133MHz, Debian, bigadv - 18846 ppd
  • ardichokeardichoke Icrontian
    edited July 2011
    Intel Dual Xeon E5506, 8 cores @ 2133MHz, Debian, SMP - 11218 ppd
  • ardichokeardichoke Icrontian
    edited July 2011
    Intel Core i5-750, 4 cores @ 2666MHz, Debian, SMP - 10993 ppd
  • _vid__vid_ alabama
    edited July 2011
    just picked up gtx 560ti, bench says 12,890.5 right now but will report back with an edit after a few days to provide more accuracy.
  • _k_k P-Town, Texas Icrontian
    edited July 2011
    just post again with core clocks when you are sure of the numbers
  • _vid__vid_ alabama
    edited July 2011
    sorry, cant edit my original post for some reason, not even running 24hours straight, 560ti netted me 13089 points, so a complete full day of folding probably be even higher than that. clocks are 900/1050/1800 which is a stock oc for this asus card.
  • _k_k P-Town, Texas Icrontian
    edited October 2011
    This is what I came up with doing testing on my new dual opteron 6134 system. The list is project, ppd, overall credit, base credit and the final set is if dynamic load balance is available for the WU.


    Ubuntu 10.04
    6062 (R1,C82,G355) 47758.53 ppd 5914 credit
    6068 (R0,C96,G437) 38633.61 ppd 4829 credit
    6079 (R0,C52,G126) 41086.09 ppd 5183 credit
    6080 (R0,C164,G90) 39928.06 ppd 5222 credit
    6940 (R0,C7,G74) 41184.88 ppd 5672 credit
    6990 (R0,C86,G38) 43957.39 ppd 5698 credit
    6992 (R0,C38,G152) 43758.94 ppd 5672 credit
    7137 (R0,C82,G131) 46714.66 ppd 5947 credit

    Ubuntu Server 11
    6060 (R0,C167,G460) 36924.16 ppd 4829 credit 481 point
    6080 (R0,C92,G90) 41017.01 ppd 5222 credit 521 point
    6970 (R0,C79,G51) 39524.20 ppd 5672 credit 552 point
    5979 (R0,C36,G301) 41184.88 ppd 5672 credit 552 point
    6946 (Ro,C134,G42) 41888.90 ppd 5672 credit 552 point
    6981 (R0,C66,G264) 42078.87 ppd 5698 credit 552 point
    6984 (R0,C88,G148) 42078.87 ppd 5698 credit 552 point
    7147 (R0,C97,G168) 43919.76 ppd 5947 credit 585 point

    Ubuntu 11
    6026 52048.53 ppd 6325 credit 517 point
    6056 37589.46 ppd 4829 credit 481 point
    6062 39732.16 ppd 4874 credit 481 point
    6068 37303.6 ppd 4564 credit 481 point
    6099 48441.42 ppd 17044 credit 1588 point
    6940 41184.88 ppd 5672 credit 552 point NO
    6942 44153.16 ppd 5672 credit 552 point YES
    6943 44153.16 ppd 5672 credit 552 point YES
    6943 42250.01 ppd 5672 credit 552 point NO
    6944 43758.94 ppd 5672 credit 552 point YES
    6950 42223.96 ppd 5033 credit 457 point YES
    6974 44153.16 ppd 5672 credit 552 point
    6980 41722.27 ppd 5698 credit 552 point
    6984 41722.27 ppd 5698 credit 552 point
    6996 44153.16 ppd 5672 credit 552 point
    7129 46714.66 ppd 5947 credit 585 point YES
    7133 58273.05 ppd 7284 credit 585 point
    7139 47579.74 ppd 5947 credit 585 point YES
    7143 47143.23 ppd 5947 credit 585 point
    7145 47143.23 ppd 5947 credit 585 point
    7148 46714.66 ppd 5947 credit 585 point YES
  • _k_k P-Town, Texas Icrontian
    Bump. Updated AMD ppd set with 6134 and 6136 from my DP rigs.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited February 2012
    Just starting, 4 raw Classic 6.34 modes running on a MSI H61M-P21 Mainboard with an i7-2600k CPU at default speed (3.4 GHz each core).

    Totall PPD so far, 3700 PPD (with FAHmon barely stable as to PPD calcs per core) approximate total. Will ramp up to SMP later, and install the MSI OC Genie II to see what OCing it can auto-generate.
  • clifford_cooleyclifford_cooley Arkansas, USA Member
    Just starting, 4 raw Classic 6.34 modes running on a MSI H61M-P21 Mainboard with an i7-2600k CPU at default speed (3.4 GHz each core).
    I think you could do better with a different chipset (P67 or Z68). Is there a reason why you chose to pair the H61 chipset with the 2600K CPU? The H61 chipset is not over-clocking chipset but the K-series CPU is known for its over-clock ability.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    think you could do better with a different chipset (P67 or Z68). Is there a reason why you chose to pair the H61 chipset with the 2600K CPU? The H61 chipset is not over-clocking chipset but the K-series CPU is known for its over-clock ability.
    Indeed I would do better with a OCing board as to production, but money is a hugely MAJOR issue. I got the mainboard I am using for $59.95 shipped, and it is a rev 3.2 board. The bottleneck on the board is BUSSING.

    Since I am deliberately letting my room hover at 80-82 Farenheit high for day today, and am only for now air cooling the case, I have in essence a CPU that is going to get about as hot as I need it to stably.

    I do not need the CPU to Speed Step slower, and affording a liquid cooling setup is months away from being something I can afford, and I will have to save for a new mainboard even. I deliberately bought a Newegg bundle to core build the system with, for about $300.00 shipped.

    My main computer is a laptop. It does not fold well or fast. It has a gen-one i7 in it.

  • TushonTushon I'm scared, Coach Alexandria, VA Icrontian
    You would do much better with SMP client now, even if you limit the number of cores it is folding on. The classic or uniprocessor client is horrible for PPD
  • pigflipperpigflipper The Forgotten Coast Icrontian
    I use the v7 SMP client and limit my i7 to two cores during the day and then let it run full tilt at night, drops my PPD a bit, but worth it to keep the system still folding even during the middle of the day.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    True, that can be done. I ramped up things my way-- I have right now seven instances of cores running on my i7's cores. Each core of the i7 2600k second-gen CPU I have will support a core running in parallel twice over. Core two runs almost all the Windows overload, of which there is little. Right now my CPU is 84 percent loaded for total load, as I run everything at 95% load, at default base of 3.4 GHz.

    I run one EXE and one Service out of each folder/directory for folding I have. I have four folders/directories. This works because each folding EXE in individually named folders (one per core) has a different name to Windows.

    Mom and I have the house AC'd during summer to 80-82 Farenheit, if the temp outside is above 85 Farenheit and/or humidity is high. So what it was today for high, is what most of summer will be like for insdie the house. Florida taught me to air cool well... :)

    John.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited February 2012
    You would do much better with SMP client now, even if you limit the number of cores it is folding on. The classic or uniprocessor client is horrible for PPD
    Um, SMP 6.34 Client x 4, no SMP switch so it runs in Classic mode for now (the docs on Folding say it will do that). Will switch after I see how many Clients I can run stably at once-- appears to be four for this CPU. Might have to leave one sans the SMP switch, so one processor pipe is free for Windows. Windows thinks it has CPUs 0-7 ( ::) ). I also have more RAM coming, I got Value grade G.Skill DDR3 1333 RAM, a 4 GB stick for $19.99 no ship fee. Expect next the RAM next Monday or Tuesday. My motherboard only has two sockets for RAM.

    John.

  • GargGarg Purveyor of Lincoln Nightmares Icrontian
    edited February 2012
    You would do much better with SMP client now, even if you limit the number of cores it is folding on. The classic or uniprocessor client is horrible for PPD
    Um, SMP 6.34 Client x 4, no SMP switch so it runs in Classic mode for now (the docs on Folding say it will do that). Will switch after I see how many Clients I can run stably at once-- appears to be four for this CPU. Might have to leave one sans the SMP switch, so one processor pipe is free for Windows. Windows thinks it has CPUs 0-7 ( ::) )...
    I know that you're trying to "ramp things up," but I think what Tushon is saying is that there is no reason to run multiple classic clients on modern processors. I do run two clients on an old P4 with HT, but that's because there's no way it could finish a SMP WU in time. Your processor could finish a SMP WU with 6 threads tied behind its back, and make more points in the process.
  • TushonTushon I'm scared, Coach Alexandria, VA Icrontian
    Whenever you change the settings, you will want to run a single SMP client with -smp X, where x is the number of cores. If you want to do some scheduling, you can do it as pigflipper does, using v7 client. Multiple clients as SMP will be a detriment.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian


    I know that you're trying to "ramp things up," but I think what Tushon is saying is that there is no reason to run multiple classic clients on modern processors. I do run two clients on an old P4 with HT, but that's because there's no way it could finish a SMP WU in time. Your processor could finish a SMP WU with 6 threads tied behind its back, and make more points in the process.
    OK, will try it that way with the v. 6.34 client for now. EACH of my clients is doing 8011's and 8001's in about 1/4 the deadline time as it is, due to sheer speed per core and subcore parallel thread processing of CPU.

    I need to find the download link and install guide for V. 7 client for Windows. I do not see one here or on Folding's main site. The Beta download link leads to install info and download of the v. 6.34 32 bit client, with instructions to run it as a -SMP or -SMP 4 (in my case, four core CPU, 2 parallel threads per core). If there is a Windows guide for V. 7 client with a download link in it, would someone link to it please???


  • GargGarg Purveyor of Lincoln Nightmares Icrontian
    I think v6 will run -smp 8. I specify N threads, but IIRC, just setting the -smp flag without specifying the number of threads should use all that are available, as it will detect that number when it starts.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited February 2012
    I think v6 will run -smp 8. I specify N threads, but IIRC, just setting the -smp flag without specifying the number of threads should use all that are available, as it will detect that number when it starts.
    Thank you muchly-- will reconfigure now.

    EDIT/ADD:
    Running an old WU, partly done. FAHMon is guessing at 6495 PPD right now. Had to use -SMP 8 to get the CPU running right. With a P8001 FAHMon is getting an estimate that has to be low. It will increase as new WUs are downloaded and run.

    Again, THANKS.

  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    PPD up to about 17,000 PPD, average for about 6 hours folding with the V. 7 client (It is doing a P8011 every 44 min average). No GPU except the built-into-CPU Intel Graphics 3000 to GPU with, so am SMP only. The CPU is pegged at 100% however, has been 98% earlier today.
  • GargGarg Purveyor of Lincoln Nightmares Icrontian
    Nice! That sounds about right given what K has listed for a 2500K on the front page. These Sandy Bridge chips pump out a lot of PPD per chip.
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    Nice! That sounds about right given what K has listed for a 2500K on the front page. These Sandy Bridge chips pump out a lot of PPD per chip.
    The PPD is stabling out at about 16,600 with a steady stream of P8011s coming. We will see what happens over a week's time, I guess. The CPU is not OC'd at all, stock base rate of 3.4 GHz folding.

  • _k_k P-Town, Texas Icrontian
    Let us get a bump on this. Since 2012 there has been a bit of hardware changes and we can greatly add to this list. I know I need to put my 7950 up here and a 4770k. There are a few new R series ATI cards in the IC group we should try and get some benchmarks from.
  • csimoncsimon Acadiana Icrontian
    I will post mine too as soon and I get more stable numbers. Is FAHControl a reliable source?
    _k
Sign In or Register to comment.