Maxwell GPUs

[billford]

As long as they're identical, you can switch them and resume work. That would eliminate the slight chance that one is positioned in a hotter part of it's case than the other.

Hadn't thought of that... thanks.

[billford]

Something of a personal best today- NVidia drivers upgraded on 3 machines without breaking an OS Maybe I've finally got the hang of it...

It's a bit early to be sure, but the 355.11 drivers seem to be producing noticeably lower PPD than the old 346.35 ones. So maybe 7im was right, and that 350-upwards drivers are better because they're more stable.

Time will tell... I'll get the "feel" of them for a few days before putting the adv flag back thus inviting some of the heavier stuff.

Picked up some useful tips in this topic, many thanks to all

[58Enfield]

CUDA-Z "http://cuda-z.sourceforge.net/" will run on debian based linux, but seems to need cuda 5.0 or higher to install. I'm still running two gtx 650 Ti's on 304.128 drivers and the installer won't work on the associated cuda 3.0. The software provides three screens of info but no capability changes.

[billford]

CUDA-Z "http://cuda-z.sourceforge.net/" will run on debian based linux, but seems to need cuda 5.0 or higher to install.

I can manage the CUDA (5.2) but the Linux is Ubuntu based (Mint 17.2). Any chance it might work?

[davidcoton]

Ubuntu is Debian based (unless I've got things horribly wrong...).

[billford]

Ubuntu is Debian based (unless I've got things horribly wrong...).

Well if you have, so has Wikipedia

Thanks, I'll look into that later on when there's half a chance I'm awake!

At the moment I'm running nvidia-smi in a 30-second "watch" loop, and at least one Core-21 WU (a P10495) is running the card very close to it's power limit... although it hasn't (yet) fallen over.

[Duce H_K_]

I just wanna say 10495 is a really power eating monster. Consumption peaks over 116% (that's limit for my mod BIOS) & gaining only 308kppd on a GTX970. Linearly thinking, PPD could be closer to this

The new core 21 tends to push the hardware harder than we're used to. If your GPU quality are very different, one might be throttling more than the other (on Windows, GPU-Z gives the value which is called PerfCap reason).

In this state, PerfCap reason showed value PWR (or smth. like that). Clock was downgrading. ASIC Quality of my GPU is 73.1%. I don't know if it's bad or good, I just consider overclock acceptable for most WUs.

Code: Select all

*********************** Log Started 2016-03-10T22:56:06Z ***********************
22:56:06:      Website: http://folding.stanford.edu/
22:56:06:    Copyright: (c) 2009-2014 Stanford University
22:56:06:       Author: Joseph Coffland <[email protected]>
22:56:06:         Args: --gpu-index=1 --opencl-index=0
22:56:06:       Config: E:/Docbase/FaH-workdir/config.xml
22:56:06:******************************** Build ********************************
22:56:06:      Version: 7.4.4
22:56:06:         Date: Mar 4 2014
22:56:06:         Time: 20:26:54
22:56:06:      SVN Rev: 4130
22:56:06:       Branch: fah/trunk/client
22:56:06:     Compiler: Intel(R) C++ MSVC 1500 mode 1200
22:56:06:      Options: /TP /nologo /EHa /Qdiag-disable:4297,4103,1786,279 /Ox -arch:SSE
22:56:06:               /QaxSSE2,SSE3,SSSE3,SSE4.1,SSE4.2 /Qopenmp /Qrestrict /MT /Qmkl
22:56:06:     Platform: win32 XP
22:56:06:         Bits: 32
22:56:06:         Mode: Release
22:56:06:******************************* System ********************************
22:56:06:          CPU: AMD A8-5600K APU with Radeon(tm) HD Graphics
22:56:06:       CPU ID: AuthenticAMD Family 21 Model 16 Stepping 1
22:56:06:         CPUs: 4
22:56:06:       Memory: 3.46GiB
22:56:06:  Free Memory: 2.58GiB
22:56:06:      Threads: WINDOWS_THREADS
22:56:06:   OS Version: 6.2
22:56:06:  Has Battery: false
22:56:06:   On Battery: false
22:56:06:   UTC Offset: 3
22:56:06:          PID: 2076
22:56:06:          CWD: E:/Docbase/FaH-workdir
22:56:06:           OS: Windows 8.1 Pro
22:56:06:      OS Arch: AMD64
22:56:06:         GPUs: 2
22:56:06:        GPU 0: ATI:5 Trinity [Radeon HD 7560D]
22:56:06:        GPU 1: NVIDIA:5 GM204 [GeForce GTX 970]
22:56:06:         CUDA: 5.2
22:56:06:  CUDA Driver: 8000
22:56:06:Win32 Service: false
22:56:06:***********************************************************************
23:06:16:<config>
23:06:16:  <!-- Folding Core -->
23:06:16:  <checkpoint v='6'/>
23:06:16:  <core-priority v='low'/>
23:06:16:
23:06:16:  <!-- Folding Slot Configuration -->
23:06:16:  <extra-core-args v='-forceasm -twait=80'/>
23:06:16:
23:06:16:  <!-- HTTP Server -->
23:06:16:  <allow v='0.0.0.0/0'/>
23:06:16:  <deny v='255.255.255.255/255.255.255.255'/>
23:06:16:
23:06:16:  <!-- Network -->
23:06:16:  <proxy v=':8080'/>
23:06:16:
23:06:16:  <!-- Remote Command Server -->
23:06:16:  <command-deny-no-pass v=''/>
23:06:16:  <command-port v='7936'/>
23:06:16:  <password v='************'/>
23:06:16:
23:06:16:  <!-- Slot Control -->
23:06:16:  <pause-on-battery v='false'/>
23:06:16:  <pause-on-start v='true'/>
23:06:16:  <power v='full'/>
23:06:16:
23:06:16:  <!-- User Information -->
23:06:16:  <passkey v='********************************'/>
23:06:16:  <team v='47191'/>
23:06:16:  <user v='Duce-HK_ALL_1P1LY4fT4YkPDEhNRFFkCQcTxbc2gr1UoB'/>
23:06:16:
23:06:16:  <!-- Work Unit Control -->
23:06:16:  <max-queue v='2'/>
23:06:16:  <next-unit-percentage v='100'/>
23:06:16:
23:06:16:  <!-- Folding Slots -->
23:06:16:  <slot id='0' type='GPU'>
23:06:16:    <gpu-index v='1'/>
23:06:16:    <client-type v='beta'/>
23:06:16:    <cuda-index v='1'/>
23:06:16:    <opencl-index v='1'/>
23:06:16:  </slot>
23:06:16:</config>
09:15:24:WU00:FS00:Connecting to 171.67.108.45:80
09:15:25:WU00:FS00:Assigned to work server 140.163.4.245
09:15:25:WU00:FS00:Requesting new work unit for slot 00: READY gpu:1:GM204 [GeForce GTX 970] from 140.163.4.245
09:15:26:WU00:FS00:Connecting to 140.163.4.245:8080
09:15:27:WU00:FS00:Downloading 5.14MiB
09:15:30:WU00:FS00:Download complete
09:15:30:WU00:FS00:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:10495 run:0 clone:40 gen:56 core:0x21 unit:0x000000448ca304f556ba62d48b9a325c
09:15:31:WU00:FS00:Starting
09:15:31:WU00:FS00:Running FahCore: C:\NO_UAC\FAHClient7/FAHCoreWrapper.exe E:/Docbase/FaH-workdir/cores/web.stanford.edu/~pande/Win32/AMD64/NVIDIA/Fermi/beta/Core_21.fah/FahCore_21.exe -dir 00 -suffix 01 -version 704 -lifeline 2076 -checkpoint 6 -gpu 1 -gpu-vendor nvidia -forceasm -twait=80
09:15:31:WU00:FS00:Started FahCore on PID 3196
09:15:31:WU00:FS00:Core PID:5024
09:15:31:WU00:FS00:FahCore 0x21 started
09:15:33:WU00:FS00:0x21:*********************** Log Started 2016-03-11T09:15:32Z ***********************
09:15:33:WU00:FS00:0x21:Project: 10495 (Run 0, Clone 40, Gen 56)
09:15:33:WU00:FS00:0x21:Unit: 0x000000448ca304f556ba62d48b9a325c
09:15:33:WU00:FS00:0x21:CPU: 0x00000000000000000000000000000000
09:15:33:WU00:FS00:0x21:Machine: 0
09:15:33:WU00:FS00:0x21:Reading tar file core.xml
09:15:33:WU00:FS00:0x21:Reading tar file system.xml
09:15:33:WU00:FS00:0x21:Reading tar file integrator.xml
09:15:33:WU00:FS00:0x21:Reading tar file state.xml
09:15:34:WU00:FS00:0x21:Digital signatures verified
09:15:34:WU00:FS00:0x21:Folding@home GPU Core21 Folding@home Core
09:15:34:WU00:FS00:0x21:Version 0.0.17
09:15:51:WU00:FS00:0x21:Completed 0 out of 5000000 steps (0%)
09:15:51:WU00:FS00:0x21:Temperature control disabled. Requirements: single Nvidia GPU, tmax must be < 110 and twait >= 900
09:18:02:WU00:FS00:0x21:Completed 50000 out of 5000000 steps (1%)
09:20:14:WU00:FS00:0x21:Completed 100000 out of 5000000 steps (2%)
09:22:25:WU00:FS00:0x21:Completed 150000 out of 5000000 steps (3%)
09:24:38:WU00:FS00:0x21:Completed 200000 out of 5000000 steps (4%)
09:26:51:WU00:FS00:0x21:Completed 250000 out of 5000000 steps (5%)
09:29:09:WU00:FS00:0x21:Completed 300000 out of 5000000 steps (6%)
09:31:25:WU00:FS00:0x21:Completed 350000 out of 5000000 steps (7%)
09:33:41:WU00:FS00:0x21:Completed 400000 out of 5000000 steps (8%)
09:35:58:WU00:FS00:0x21:Completed 450000 out of 5000000 steps (9%)
09:38:14:WU00:FS00:0x21:Completed 500000 out of 5000000 steps (10%)
09:40:35:WU00:FS00:0x21:Completed 550000 out of 5000000 steps (11%)
09:42:52:WU00:FS00:0x21:Completed 600000 out of 5000000 steps (12%)
09:45:09:WU00:FS00:0x21:Completed 650000 out of 5000000 steps (13%)
09:47:26:WU00:FS00:0x21:Completed 700000 out of 5000000 steps (14%)
09:49:43:WU00:FS00:0x21:Completed 750000 out of 5000000 steps (15%)
09:52:05:WU00:FS00:0x21:Completed 800000 out of 5000000 steps (16%)
09:54:22:WU00:FS00:0x21:Completed 850000 out of 5000000 steps (17%)
09:56:41:WU00:FS00:0x21:Completed 900000 out of 5000000 steps (18%)
09:58:59:WU00:FS00:0x21:Completed 950000 out of 5000000 steps (19%)
10:01:16:WU00:FS00:0x21:Completed 1000000 out of 5000000 steps (20%)
10:03:39:WU00:FS00:0x21:Completed 1050000 out of 5000000 steps (21%)
10:05:57:WU00:FS00:0x21:Completed 1100000 out of 5000000 steps (22%)
10:08:14:WU00:FS00:0x21:Completed 1150000 out of 5000000 steps (23%)
10:10:30:WU00:FS00:0x21:Completed 1200000 out of 5000000 steps (24%)
10:12:47:WU00:FS00:0x21:Completed 1250000 out of 5000000 steps (25%)
10:15:07:WU00:FS00:0x21:Completed 1300000 out of 5000000 steps (26%)
10:17:24:WU00:FS00:0x21:Completed 1350000 out of 5000000 steps (27%)
10:19:40:WU00:FS00:0x21:Completed 1400000 out of 5000000 steps (28%)
10:21:56:WU00:FS00:0x21:Completed 1450000 out of 5000000 steps (29%)
10:24:12:WU00:FS00:0x21:Completed 1500000 out of 5000000 steps (30%)
10:26:32:WU00:FS00:0x21:Completed 1550000 out of 5000000 steps (31%)
10:28:48:WU00:FS00:0x21:Completed 1600000 out of 5000000 steps (32%)
10:31:04:WU00:FS00:0x21:Completed 1650000 out of 5000000 steps (33%)
10:33:20:WU00:FS00:0x21:Completed 1700000 out of 5000000 steps (34%)
* * * * 10:34:35:WU00:FS00:0x21:Bad State detected... attempting to resume from last good checkpoint. Is your system overclocked? * * * *
10:36:51:WU00:FS00:0x21:Completed 1550000 out of 5000000 steps (31%)
10:39:07:WU00:FS00:0x21:Completed 1600000 out of 5000000 steps (32%)
10:41:24:WU00:FS00:0x21:Completed 1650000 out of 5000000 steps (33%)
10:43:40:WU00:FS00:0x21:Completed 1700000 out of 5000000 steps (34%)
10:45:56:WU00:FS00:0x21:Completed 1750000 out of 5000000 steps (35%)
10:48:19:WU00:FS00:0x21:Completed 1800000 out of 5000000 steps (36%)
10:50:36:WU00:FS00:0x21:Completed 1850000 out of 5000000 steps (37%)
10:52:54:WU00:FS00:0x21:Completed 1900000 out of 5000000 steps (38%)
10:55:12:WU00:FS00:0x21:Completed 1950000 out of 5000000 steps (39%)
10:57:30:WU00:FS00:0x21:Completed 2000000 out of 5000000 steps (40%)
10:59:53:WU00:FS00:0x21:Completed 2050000 out of 5000000 steps (41%)
11:02:10:WU00:FS00:0x21:Completed 2100000 out of 5000000 steps (42%)
11:04:28:WU00:FS00:0x21:Completed 2150000 out of 5000000 steps (43%)
11:06:45:WU00:FS00:0x21:Completed 2200000 out of 5000000 steps (44%)
11:09:03:WU00:FS00:0x21:Completed 2250000 out of 5000000 steps (45%)
11:11:26:WU00:FS00:0x21:Completed 2300000 out of 5000000 steps (46%)
11:13:43:WU00:FS00:0x21:Completed 2350000 out of 5000000 steps (47%)
11:16:01:WU00:FS00:0x21:Completed 2400000 out of 5000000 steps (48%)
11:18:19:WU00:FS00:0x21:Completed 2450000 out of 5000000 steps (49%)
11:20:36:WU00:FS00:0x21:Completed 2500000 out of 5000000 steps (50%)
11:22:58:WU00:FS00:0x21:Completed 2550000 out of 5000000 steps (51%)
11:25:16:WU00:FS00:0x21:Completed 2600000 out of 5000000 steps (52%)
11:27:34:WU00:FS00:0x21:Completed 2650000 out of 5000000 steps (53%)
11:29:52:WU00:FS00:0x21:Completed 2700000 out of 5000000 steps (54%)
11:32:10:WU00:FS00:0x21:Completed 2750000 out of 5000000 steps (55%)
11:34:32:WU00:FS00:0x21:Completed 2800000 out of 5000000 steps (56%)
11:36:50:WU00:FS00:0x21:Completed 2850000 out of 5000000 steps (57%)
11:39:08:WU00:FS00:0x21:Completed 2900000 out of 5000000 steps (58%)
11:41:26:WU00:FS00:0x21:Completed 2950000 out of 5000000 steps (59%)
11:43:44:WU00:FS00:0x21:Completed 3000000 out of 5000000 steps (60%)
11:46:07:WU00:FS00:0x21:Completed 3050000 out of 5000000 steps (61%)
11:48:26:WU00:FS00:0x21:Completed 3100000 out of 5000000 steps (62%)
11:50:43:WU00:FS00:0x21:Completed 3150000 out of 5000000 steps (63%)
11:53:02:WU00:FS00:0x21:Completed 3200000 out of 5000000 steps (64%)
11:55:19:WU00:FS00:0x21:Completed 3250000 out of 5000000 steps (65%)
11:57:42:WU00:FS00:0x21:Completed 3300000 out of 5000000 steps (66%)
12:00:00:WU00:FS00:0x21:Completed 3350000 out of 5000000 steps (67%)
12:02:19:WU00:FS00:0x21:Completed 3400000 out of 5000000 steps (68%)
12:04:37:WU00:FS00:0x21:Completed 3450000 out of 5000000 steps (69%)
12:06:55:WU00:FS00:0x21:Completed 3500000 out of 5000000 steps (70%)
12:09:17:WU00:FS00:0x21:Completed 3550000 out of 5000000 steps (71%)
12:11:35:WU00:FS00:0x21:Completed 3600000 out of 5000000 steps (72%)
12:13:53:WU00:FS00:0x21:Completed 3650000 out of 5000000 steps (73%)
12:16:11:WU00:FS00:0x21:Completed 3700000 out of 5000000 steps (74%)
12:18:29:WU00:FS00:0x21:Completed 3750000 out of 5000000 steps (75%)
12:20:52:WU00:FS00:0x21:Completed 3800000 out of 5000000 steps (76%)
12:23:09:WU00:FS00:0x21:Completed 3850000 out of 5000000 steps (77%)
12:25:27:WU00:FS00:0x21:Completed 3900000 out of 5000000 steps (78%)
12:27:45:WU00:FS00:0x21:Completed 3950000 out of 5000000 steps (79%)
12:30:03:WU00:FS00:0x21:Completed 4000000 out of 5000000 steps (80%)
12:32:25:WU00:FS00:0x21:Completed 4050000 out of 5000000 steps (81%)
12:34:42:WU00:FS00:0x21:Completed 4100000 out of 5000000 steps (82%)
12:37:00:WU00:FS00:0x21:Completed 4150000 out of 5000000 steps (83%)
12:39:18:WU00:FS00:0x21:Completed 4200000 out of 5000000 steps (84%)
12:41:36:WU00:FS00:0x21:Completed 4250000 out of 5000000 steps (85%)
12:43:57:WU00:FS00:0x21:Completed 4300000 out of 5000000 steps (86%)
12:46:15:WU00:FS00:0x21:Completed 4350000 out of 5000000 steps (87%)
12:48:33:WU00:FS00:0x21:Completed 4400000 out of 5000000 steps (88%)
12:50:51:WU00:FS00:0x21:Completed 4450000 out of 5000000 steps (89%)
12:53:09:WU00:FS00:0x21:Completed 4500000 out of 5000000 steps (90%)
12:55:31:WU00:FS00:0x21:Completed 4550000 out of 5000000 steps (91%)
12:57:49:WU00:FS00:0x21:Completed 4600000 out of 5000000 steps (92%)
13:00:07:WU00:FS00:0x21:Completed 4650000 out of 5000000 steps (93%)
13:02:26:WU00:FS00:0x21:Completed 4700000 out of 5000000 steps (94%)
13:04:44:WU00:FS00:0x21:Completed 4750000 out of 5000000 steps (95%)
13:07:06:WU00:FS00:0x21:Completed 4800000 out of 5000000 steps (96%)
13:09:24:WU00:FS00:0x21:Completed 4850000 out of 5000000 steps (97%)
13:11:42:WU00:FS00:0x21:Completed 4900000 out of 5000000 steps (98%)
13:14:00:WU00:FS00:0x21:Completed 4950000 out of 5000000 steps (99%)
13:16:18:WU00:FS00:0x21:Completed 5000000 out of 5000000 steps (100%)
13:16:23:WU00:FS00:0x21:Saving result file logfile_01.txt
* * * * 13:16:23:WU00:FS00:0x21:Saving result file badstate-0.xml * * * *
13:16:29:WU00:FS00:0x21:Saving result file checkpointState.xml
13:16:32:WU00:FS00:0x21:Saving result file checkpt.crc
13:16:32:WU00:FS00:0x21:Saving result file log.txt
13:16:32:WU00:FS00:0x21:Saving result file positions.xtc
13:16:34:WU00:FS00:0x21:Folding@home Core Shutdown: FINISHED_UNIT
13:16:36:WU00:FS00:FahCore returned: FINISHED_UNIT (100 = 0x64)
13:16:36:WU00:FS00:Sending unit results: id:00 state:SEND error:NO_ERROR project:10495 run:0 clone:40 gen:56 core:0x21 unit:0x000000448ca304f556ba62d48b9a325c
13:16:36:WU00:FS00:Uploading 8.76MiB to 140.163.4.245
13:16:36:WU00:FS00:Connecting to 140.163.4.245:8080
13:16:45:WU00:FS00:Upload 40.68%
13:16:51:WU00:FS00:Upload 82.07%
13:17:04:WU00:FS00:Upload complete
13:17:04:WU00:FS00:Server responded WORK_ACK (400)
13:17:04:WU00:FS00:Final credit estimate, 52810.00 points

I'm glad new core was wise enough to ask me 'Did you play with overclocking?' Honestly I did.

Also, core 21 projects are often bigger (number of atoms) and the sanity checks take more time. Those sanity checks are performed on the CPU, so different CPUs might also be a reason that exacerbate the differences.

That's right. As i watched through recent time, smaller 0x21 core WUs eat 2 threads (36-50% of a quad), but the higher the atoms number - the higher cpu load on checkpoints.

[bruce]

... but the higher the atoms number - the higher cpu load on checkpoints.

That's to be expected. The calculations done on the CPU probably increase with the square of the atom count.