|
本帖最后由 金鹏 于 2020-1-5 16:08 编辑
PPD和温度还是一如既往地高,PPD貌似和以前11733包差不多
A卡加advanced或beta参数可以跑了,可以跑了,可以跑了
GPU CORE22 0.0.2 coming to ADVANCED[size=1.3em]The updated GPU Core 22 is here! This is based on OpenMM 7.4.1 - we're starting with the OpenCL version first. Stability has drastically improved from 0.0.1 -- I'm seeing 0.3% failure rate in BETA.
The deadline and credit were adjusted to increase the PPDs by the performance speedup / increase in science produced (speedup is ~20% on 1080/1080Ti benchmark machine). Please watch out for temperature and wattage increases, and consider reducing OC.
Project stats:
11737 (DHFR benchmark)
core: Core22 0.0.2
# atoms: 23558
credit: 7498
k-factor: 0.75
timeout: 5.8d
deadline: 8.2d
precision: mixed
It will take another day or two before I make it active -- this is an advance notice -- and I will go in small batches of letting a few hundred WUs out and then switching off ADVANCED for a bit and so on -- so that if there's problems they'll be contained to a small amount of donors, and if you have failures other projects will be downloaded instead.
Thank you all! - 08:48:48:WU00:FS01:Starting
- 08:48:48:WU00:FS01:Running FahCore: C:\FAHClient/FAHCoreWrapper.exe C:\FAHClient\FAHClient\cores/cores.foldingathome.org/v7/win/64bit/Core_22.fah/FahCore_22.exe -dir 00 -suffix 01 -version 705 -lifeline 7604 -checkpoint 3 -gpu-vendor nvidia -opencl-platform 0 -opencl-device 0 -cuda-device 0 -gpu 0
- 08:48:48:WU00:FS01:Started FahCore on PID 15696
- 08:48:48:Started thread 12 on PID 7604
- 08:48:48:WU00:FS01:Core PID:14356
- 08:48:48:WU00:FS01:FahCore 0x22 started
- 08:48:49:WU00:FS01:0x22:*********************** Log Started 2019-12-14T08:48:48Z ***********************
- 08:48:49:WU00:FS01:0x22:*************************** Core22 Folding@home Core ***************************
- 08:48:49:WU00:FS01:0x22: Type: 0x22
- 08:48:49:WU00:FS01:0x22: Core: Core22
- 08:48:49:WU00:FS01:0x22: Website: https://foldingathome.org/
- 08:48:49:WU00:FS01:0x22: Copyright: (c) 2009-2018 foldingathome.org
- 08:48:49:WU00:FS01:0x22: Author: John Chodera <[email protected]> and Rafal Wiewiora
- 08:48:49:WU00:FS01:0x22: <[email protected]>
- 08:48:49:WU00:FS01:0x22: Args: -dir 00 -suffix 01 -version 705 -lifeline 15696 -checkpoint 3
- 08:48:49:WU00:FS01:0x22: -gpu-vendor nvidia -opencl-platform 0 -opencl-device 0 -cuda-device
- 08:48:49:WU00:FS01:0x22: 0 -gpu 0
- 08:48:49:WU00:FS01:0x22: Config: <none>
- 08:48:49:WU00:FS01:0x22:************************************ Build *************************************
- 08:48:49:WU00:FS01:0x22: Version: 0.0.2
- 08:48:49:WU00:FS01:0x22: Date: Dec 6 2019
- 08:48:49:WU00:FS01:0x22: Time: 21:30:31
- 08:48:49:WU00:FS01:0x22: Repository: Git
- 08:48:49:WU00:FS01:0x22: Revision: abeb39247cc72df5af0f63723edafadb23d5dfbe
- 08:48:49:WU00:FS01:0x22: Branch: HEAD
- 08:48:49:WU00:FS01:0x22: Compiler: Visual C++ 2008
- 08:48:49:WU00:FS01:0x22: Options: /TP /nologo /EHa /wd4297 /wd4103 /Ox /MT
- 08:48:49:WU00:FS01:0x22: Platform: win32 10
- 08:48:49:WU00:FS01:0x22: Bits: 64
- 08:48:49:WU00:FS01:0x22: Mode: Release
- 08:48:49:WU00:FS01:0x22:************************************ System ************************************
- 08:48:49:WU00:FS01:0x22: CPU: Intel(R) Xeon(R) CPU E3-1280 v3 @ 3.60GHz
- 08:48:49:WU00:FS01:0x22: CPU ID: GenuineIntel Family 6 Model 60 Stepping 3
- 08:48:49:WU00:FS01:0x22: CPUs: 8
- 08:48:49:WU00:FS01:0x22: Memory: 9.95GiB
- 08:48:49:WU00:FS01:0x22:Free Memory: 5.18GiB
- 08:48:49:WU00:FS01:0x22: Threads: WINDOWS_THREADS
- 08:48:49:WU00:FS01:0x22: OS Version: 6.2
- 08:48:49:WU00:FS01:0x22:Has Battery: false
- 08:48:49:WU00:FS01:0x22: On Battery: false
- 08:48:49:WU00:FS01:0x22: UTC Offset: 8
- 08:48:49:WU00:FS01:0x22: PID: 14356
- 08:48:49:WU00:FS01:0x22: CWD: C:\FAHClient\FAHClient\work
- 08:48:49:WU00:FS01:0x22: OS: Windows 10 Pro
- 08:48:49:WU00:FS01:0x22: OS Arch: AMD64
- 08:48:49:WU00:FS01:0x22:********************************************************************************
- 08:48:49:WU00:FS01:0x22:Project: 11737 (Run 0, Clone 48, Gen 12)
- 08:48:49:WU00:FS01:0x22:Unit: 0x0000000c8ca304f15dead60d432b7e6e
- 08:48:49:WU00:FS01:0x22:Reading tar file core.xml
- 08:48:49:WU00:FS01:0x22:Reading tar file integrator.xml
- 08:48:49:WU00:FS01:0x22:Reading tar file state.xml
- 08:48:49:WU00:FS01:0x22:Reading tar file system.xml
- 08:48:49:WU00:FS01:0x22:Digital signatures verified
- 08:48:49:WU00:FS01:0x22:Folding@home GPU Core22 Folding@home Core
- 08:48:49:WU00:FS01:0x22:Version 0.0.2
- 08:48:52:WU00:FS01:0x22:Completed 0 out of 5000000 steps (0%)
- 08:48:52:WU00:FS01:0x22:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
- 08:48:54:WU01:FS01:Upload 63.06%
- 08:49:21:WU01:FS01:Upload complete
- 08:49:21:WU01:FS01:Server responded WORK_ACK (400)
- 08:49:21:WU01:FS01:Final credit estimate, 185349.00 points
- 08:49:21:WU01:FS01:Cleaning up
- 08:49:36:WU00:FS01:0x22:Completed 50000 out of 5000000 steps (1%)
- 08:50:20:WU00:FS01:0x22:Completed 100000 out of 5000000 steps (2%)
- 08:51:03:WU00:FS01:0x22:Completed 150000 out of 5000000 steps (3%)
- 08:51:47:WU00:FS01:0x22:Completed 200000 out of 5000000 steps (4%)
- 08:52:31:WU00:FS01:0x22:Completed 250000 out of 5000000 steps (5%)
复制代码
=====================================================
据@yimu35 兄弟的2050MHZ下1080TI实测,PPD提升20+%
========================
从1080跑CORE22内核来看提升PPD有20+%,显卡温度和功耗相对增大,最明显的MC使用率达到近40%,说明显卡效率提升很大
================================================
从2025MHZ下的1080TI来看PPD提升约25+%,功耗和温度大升,公版1080TI风扇从原来的70%转速70-75度飙升到85%转速的80度(自定义风扇策略)
GPU CORE22 (OpenMM 7.2, Win/Lin, OpenCL) going to ADVANCED[size=1.3em]The new GPU Core 22 is here! This is based on OpenMM 7.2 - we're starting with the OpenCL version first, with CUDA joining in one of the next few versions. Stability is as good as Core 21 in beta testing - there were reports of AMD GPU underutilization that we would like to hear more about if you notice it.
The deadline and credit were adjusted to increase the PPDs by the performance speedup / increase in science produced (speedup is ~20% on 1080/1080Ti benchmark machine). Please watch out for temperature and wattage increases, and consider reducing OC.
Project stats:
11733 (DHFR benchmark)
core: Core22 0.0.1
# atoms: 23558
credit: 7498
k-factor: 0.75
timeout: 5.8d
deadline: 8.2d
precision: mixed
Thank you all! CORE22 (OpenMM 7.2, Win/Lin, OpenCL) going to BETA[size=1.3em]Hi!
The GPU Core22 (based on OpenMM 7.2) is ready for testing. We're now on version 0.0.1 with the PME stream disabled to increase stability. This is OpenCL only first, with CUDA coming in one of the next versions.
You might still observe bad states sometimes that should mostly recover. Please report failures, and bad states too if you can. We're now getting back nice science logs with this core so I can diagnose problems quickly.
We had one report of consistent failures on PCIe 1x risers in internal testing - please let me know if you're using risers here, and watch out for problems.
The Core22/Core21 speed-up is ~ 23% average (780/980/1080/1080Ti). On my new benchmark 1080/1080Ti machine I have 18% speed-up. The PPD reward is there - to start with, the deadlines are decreasing by the 18% to 5.8/8.2, credit staying the same - from the Core21 DHFR benchmark project. We'll see what the PPD increases are like in testing, and consider revising the adjustment to increase the PPDs further later - let me know what you think. Titan Xp: ("Typical" 1.2 Million, though variable depending on simulation size).
Avg. Time / Frame : 00:00:51 - 1,191,094 PPD
GTX 1070: ("Typical" 600k, though variable depending on simulation size).
Avg. Time / Frame : 00:01:22 - 584,225 PPD
Radeon VII: ("Typical 1.4 Million, though HIGHLY variable depending on simulation size).
Avg. Time / Frame : 00:01:10 - 740,720 PPD
RX480: ("Typical 325k, though HIGHLY variable depending on simulation size).
Avg. Time / Frame : 00:02:17 - 270,533 PPD
OC'ed GTX 980: ("Typical" 450k, though variable depending on simulation size).
Avg. Time / Frame : 00:01:35 - 468,507 PPD
GTX 980: ("Typical" 400k, though variable depending on simulation size).
Avg. Time / Frame : 00:01:46 - 397,505 PPD
|
评分
-
查看全部评分
|