All we can say for certain is that there was a change in behaviour after the 10.18.10.3621 driver, which produces a high validation rate. Quite what goes wrong - what changes in the result file - I'm afraid I can't say: that might be easier for you or Bikeman to extract from the returned results.
For anyone needing the last driver that works, I've looked through my downloads (I checked the readme), and my download history (I use Getright),
the original (direct) download url for the 10.18.10.3621 driver for 64bit Windows is:
This document makes references to products developed by Intel. There are some restrictions on how these products may be used, and what information may be disclosed to others. Please read the Disclaimer section and contact your Intel field representative if you would like more information.
III. Disclaimer
Intel is making no claims of usability, efficacy or warranty. The INTEL SOFTWARE LICENSE AGREEMENT contained herein completely defines the license and use of this software.
This document contains information on products in the design phase of development. The information here is subject to change without notice. Do not finalize a design with this information.
IV. Important Note
Your computer is configured to run the WinSAT* test, which determines whether the computer can support the Windows Aero* desktop theme. The test will run after this driver is installed and the computer is restarted. If your computer can support it, the Windows Aero desktop theme will automatically be enabled.
This test will take approximately 10 seconds.
*Other names and brands may be claimed as the property of others.
Copyright (C) 2014 Intel Corporation. All rights reserved.
My fleet of Android and Linux Arm hosts have five invalid tasks, and a number of inconclusives between them where they have been matched with a pair of intel_gpu hosts: http://einsteinathome.org/account/127061/computers
Thanks, Claggy. I've checked your invalids and 3 out of 4 (which are still there) showed 2 Haswell hosts validating each other against your system. That's a strong hint that whatever the new Intel drivers are doing is not happening under Linux / Android.
The 4th WU is a validation nightmare, but at least contains an example of a Haswell validating against the NEON app. That Haswell might simply use the older driver, though.
The 4th WU is a validation nightmare, but at least contains an example of a Haswell validating against the NEON app. That Haswell might simply use the older driver, though.
That one should be good, judging by Bikeman's remarks about the problems we had at the other end of the driver range when the application was first launched.
The 4th WU is a validation nightmare, but at least contains an example of a Haswell validating against the NEON app. That Haswell might simply use the older driver, though.
There is a way of finding out (the present driver version), while you can't see it by looking at the host details page, you can by looking at the last contact as long as they ask for work:
2014-12-17 07:01:19.9178 [PID=23748] [version] Checking plan class 'opencl-intel_gpu'
2014-12-17 07:01:19.9179 [PID=23748] [version] parsed project prefs setting 'gpu_util_brp': 0.000000
2014-12-17 07:01:19.9179 [PID=23748] [version] OpenCL driver version: 10.18.10.3977; platform version: OpenCL 1.2; device version: OpenCL 1.2
2014-12-17 07:01:19.9179 [PID=23748] [version] Peak flops supplied: 6.4e+10
2014-12-17 07:01:19.9179 [PID=23748] [version] plan class ok
Claggy
Ouch. That one's newer than the downloads I found on Monday, and I can't find it on Intel's site - and it still doesn't seem to be working. I hope it's not an OEM special.
The last host isn't completing any intel_gpu Wu's, some end in maximum time exceeded, maybe some a more allowance is needed, But it's a huge difference between the valid result of 904.63secs (on the other host), and the 27,726.57secs some of its results take,
looking at the stderr.txt shows no checkpointing recorded in any of those long running Wu's, some throw an Access Violation error:
There are discussions about Intel driver version and cpu's with Intel HD graphics.
Currently i'm installing a low end system.
28.12.2014 20:05:14 | | OpenCL: Intel GPU 0: Intel(R) HD Graphics (driver version 10.18.10.3408, device version OpenCL 1.2, 1496MB, 1496MB available, 6 GFLOPS peak)
28.12.2014 20:05:14 | | OpenCL CPU: Intel(R) Celeron(R) CPU J1900 @ 1.99GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 3.0.1.10878, device version OpenCL 1.2 (Build 76413))
28.12.2014 20:05:14 | | Host name: rocky1
Driver is from the distribution disc. Have not yet checked for updates.
When installation is finished and all updates done i will download some intel wu's so that everyone who is interested can check performance and reliability.
RE: All we can say for
)
For anyone needing the last driver that works, I've looked through my downloads (I checked the readme), and my download history (I use Getright),
the original (direct) download url for the 10.18.10.3621 driver for 64bit Windows is:
http://downloadmirror.intel.com/23885/a08/win64_153322.zip
Claggy
RE: My fleet of Android and
)
Thanks, Claggy. I've checked your invalids and 3 out of 4 (which are still there) showed 2 Haswell hosts validating each other against your system. That's a strong hint that whatever the new Intel drivers are doing is not happening under Linux / Android.
The 4th WU is a validation nightmare, but at least contains an example of a Haswell validating against the NEON app. That Haswell might simply use the older driver, though.
MrS
Scanning for our furry friends since Jan 2002
RE: The 4th WU is a
)
Actually, it uses an even older driver:
That one should be good, judging by Bikeman's remarks about the problems we had at the other end of the driver range when the application was first launched.
RE: The 4th WU is a
)
There is a way of finding out (the present driver version), while you can't see it by looking at the host details page, you can by looking at the last contact as long as they ask for work:
Host 7454946:
http://einstein.phys.uwm.edu/host_sched_logs/7454/7454946
Host 7536812:
http://einstein.phys.uwm.edu/host_sched_logs/7536/7536812
Host 8362996
http://einstein.phys.uwm.edu/host_sched_logs/8362/8362996
Host 11672358
http://einstein.phys.uwm.edu/host_sched_logs/11672/11672358
Host 10823173:
http://einstein.phys.uwm.edu/host_sched_logs/10823/10823173
Host 11674682:
http://einstein.phys.uwm.edu/host_sched_logs/11674/11674682
Claggy
Good to know.. thanks,
)
Good to know.. thanks, guys!
MrS
Scanning for our furry friends since Jan 2002
Host 11672358 has asked for
)
Host 11672358 has asked for work:
Claggy
RE: Host 11672358 has asked
)
Ouch. That one's newer than the downloads I found on Monday, and I can't find it on Intel's site - and it still doesn't seem to be working. I hope it's not an OEM special.
The last host isn't
)
The last host isn't completing any intel_gpu Wu's, some end in maximum time exceeded, maybe some a more allowance is needed, But it's a huge difference between the valid result of 904.63secs (on the other host), and the 27,726.57secs some of its results take,
looking at the stderr.txt shows no checkpointing recorded in any of those long running Wu's, some throw an Access Violation error:
All tasks for computer 7536812
While it hasn't asked for work here, it is attached to other projects, a stderr.txt from a Seti v7 result shows:
http://setiathome.berkeley.edu/show_host_detail.php?hostid=7026234
Claggy
There are discussions about
)
There are discussions about Intel driver version and cpu's with Intel HD graphics.
Currently i'm installing a low end system.
28.12.2014 20:05:14 | | OpenCL: Intel GPU 0: Intel(R) HD Graphics (driver version 10.18.10.3408, device version OpenCL 1.2, 1496MB, 1496MB available, 6 GFLOPS peak)
28.12.2014 20:05:14 | | OpenCL CPU: Intel(R) Celeron(R) CPU J1900 @ 1.99GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 3.0.1.10878, device version OpenCL 1.2 (Build 76413))
28.12.2014 20:05:14 | | Host name: rocky1
Driver is from the distribution disc. Have not yet checked for updates.
When installation is finished and all updates done i will download some intel wu's so that everyone who is interested can check performance and reliability.
Alexander
First wu's are finished. They
)
First wu's are finished. They validate against neon wu's.
One wu was distributed to a third pc. The wu cruched by an i5 HD4600 was marked invalid then.
The user has a lot of invalids (345) and some hundred in the queue.
http://einsteinathome.org/host/11706755/tasks&offset=0&show_names=1&state=4&appid=19