Hello Matt,

here my findings. 0.59 runs without failures and validates well. (Win7 64 bit, 2x5850, Catalyst 11.3). It still wants to use a full core (then GPU-usage about 99 % and CPU-usage 100 %). With full load on all cores with BOINC CPU-Projects 0.59 shows 0 % CPU-usage, but the GPU-usage drops to 62-66 %, which results in a longer runtime of 50 % more. It scales well with the GPU-usage.

Cheers Nowi

---

This is embarassing. I think the project would be better off if they just restored last week's version and our administrators backed off, took a few deep breaths and did some real software engineering and project planning.

If this were a release from a for-profit software company I would agree with you. Considering this is a volunteer science project run on a very tight budget I have to disagree.

There was an initial unplanned oops which took the project down and was going to necessitate a few days to recover. What better time to throw in the upcoming changes that might mess things up and take the project down for a few days? It also saves a lot of time and money avoiding all that pesky beta testing, scheduled roll outs and downtime. People tend to be more productive when they're under the gun. ;-)

It's a valid point. No matter if volunteer or not, this either slows the project down massively or cuts into resources for other projects if you decide to dedicate cores to keep the speed. On top, missing optimization increases losses by load/unload times of the GPUs. This means less work being done for MW and for other projects. No matter if you call this embarassing or not, it's far away from being any progress.

I'm finding exactly the same as [Boinc.at]Nowi and my rigs are running 2 x CPU at Seti.

HD5850, Win 7 64 SP1, BM 6.10.58, Cat 11.2, Pent D 3.2 HT
0.57 CPU 25% GPU 75%
0.59 CPU 0% GPU 65%

HD4870, XP32 SP3, BM 6.10.60, Cat 11.3, Pent D 2.8 HT
0.57 CPU 20% GPU 83%
0.59 CPU 0% GPU 65%

All we need now is a means of running 2 workunits at a time, adjust the credit given, and were back in business!!!

Cheers

---

Don't drink water, that's the stuff that rusts pipes

I'm having major problems on one machine
P4 3.8 HD4850 XPSP3

All WUs error out after 6 seconds with the following message.

<core_client_version>6.10.56</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
<search_application> milkywayathome_client separation 0.59 Windows x86 double CAL++ </search_application>
Found 1 CAL devices
Chose device 0

Device target: CAL_TARGET_770
Revision: 2
CAL Version: 1.4.838
Engine clock: 625 Mhz
Memory clock: 993 Mhz
GPU RAM: 512
Wavefront size: 64
Double precision: CAL_TRUE
Compute shader: CAL_TRUE
Number SIMD: 10
Number shader engines: 1
Pitch alignment: 256
Surface alignment: 256
Max size 2D: { 8192, 8192 }

Failed to map resource: Operational error (CAL_RESULT_ERROR)
Failed to release CAL resource
Failed to zero output buffer: The resource in question is still in use (CAL_RESULT_ERROR)
Failed to create output buffer: The resource in question is still in use (CAL_RESULT_ERROR)
Failed to map resource: Operational error (CAL_RESULT_ERROR)
Failed to release CAL resource
Failed to zero output buffer: The resource in question is still in use (CAL_RESULT_ERROR)
Failed to create out streams buffer: The resource in question is still in use (CAL_RESULT_ERROR)
Failed to create buffers: No error (CAL_RESULT_ERROR)
Failed to release CAL resource: A handle parameter is invalid (CAL_RESULT_BAD_HANDLE)
Failed to release buffers: No error (CAL_RESULT_BAD_HANDLE)
Integral 0 time = 3.476377 s
Failed to calculate integral 0
10:48:13 (2712): called boinc_finish

I've restarted Boinc and rebooted but every one is the same.

---

Is the new ATi application using OpenCL?

My old Prescott 3.2GHz P4 w HT with the HD3850 AGP and 11.3 drivers is now working with 0.59 on 32bit XP (thank you). Interestingly its CPU load is pretty much 0.

On my Q9450 running 64 bit Win 7 with the HD5970 with 0.59 it is still taking pretty much a whole cpu core per GPU wu which is slowing down the CPU BOINC projects drastically. (ps I just noticed my machine downgraded to 10.11 drivers when I had a crash while I was at work and my family rebooted it, I'll try the 11.3 drivers again)

@ Bigred - Are you running cat 11.3 on XP? if not you need to.

@GG - Great! Is that the hotfix vrsion?

---

Don't drink water, that's the stuff that rusts pipes

@GG - Great! Is that the hotfix vrsion?

Sure is! :)

A shutdown, unplug and restart seems to have solved the problem.

Hello,
ar the new wus larger than the old 0.2X app WUs were? My setup used to do the old WUs in about 1:15, now i need over 3 minutes for 2 WUs (i'm running 2 WUs, because the gpu utilisation is very low with the new app, only about 80%; used to be 98% to 99% with the old app).
Would be great, if you could raise the gpu utilisation in further releases up to 99%, as users could still have a snappy system with higher responsiveness factor.

Well I've now discovered why my ATI cards cannot compete in output with you guys!

On each rig I've been running Seti on the CPU's, and of course getting 65% GPU utilsation. I've now suspended all CPU work on the rigs and the cards now do 99% utilisation and workunits in half the time :-)

It probably doesn't matter on quads, but all my rigs run Pent 4 D's with HT. Under the old Milkyway I could run 2 x Seti at a time plus the GPU at 95% on Milkyway. Now it looks like I can't :-(

So the outcome is I will still have to keep a couple of rigs running Collatz, and maybe DNETC if it ever comes back, just so I can run Seti......

Unless anyone knows any different ?

@ Bigred & GG - glad it's sorted for you both :-))

---

Don't drink water, that's the stuff that rusts pipes

It does matter on quads - there is a high CPU utilisation. Where no CPU WUs are run, then times are at or about 0.23. Where a CPU WU is run, times are 25% or so slower.

So for those who want to run CPU apps, its currently a no no, and for the moment GPU app is elsewhere.

Its a new app, written from scratch, bound to be teething problems, but its looking as though it will be fine long run when the CPU utilisation is resolved. Its a pity the app was not run as a separate choice in preferences as a Public Beta, as at present being Prime Time app there's a lot of pressure to get it to "perform".

It'll happen, meanwhile for those not wanting to run a CPU WU, all appears fine.

Regards
Zy

Thanks ZY, you've confirmed what I had more or less surmised for myself, that CPU projects are no go at the moment on MW GPU rigs. Oh well, I'll just have to put up with it for the time being I suppose. I hope the credit we are going to be getting is gonna reflect that though :-))

---

Don't drink water, that's the stuff that rusts pipes

Q9450 Win 7 64bit. HD5970 w 11.4 early preview drivers. MW 0.59 still taking a whole CPU core per GPU wu.

No high CPU utilisation with 0.59 for me, yet the GPU usage is at about 80% when crunching on the main processor and 100% when only the ATI app is running. Hope you'll fix that soon :)

.... I hope the credit we are going to be getting is gonna reflect that though :-))

That creeking noise in the background is the sound of Pandora's Box being opened rofl :)

Regards
Zy

I'm finding exactly the same as [Boinc.at]Nowi and my rigs are running 2 x CPU at Seti.

All we need now is a means of running 2 workunits at a time, adjust the credit given, and were back in business!!!

Cheers

I am running two workunits and it does increase the GPU usage up to 100% and the CPU usage is still at 0%

I don't get the high CPU usage either it bounces up to 13-23% (1 core=25%) for about three or four seconds at the end but that is like 0.23.

I am running two workunits and it does increase the GPU usage up to 100% and the CPU usage is still at 0%

How are you running two workunits concurrently ?

---

Don't drink water, that's the stuff that rusts pipes

It does matter on quads - there is a high CPU utilisation. Where no CPU WUs are run, then times are at or about 0.23. Where a CPU WU is run, times are 25% or so slower.

So for those who want to run CPU apps, its currently a no no, and for the moment GPU app is elsewhere.

Its a new app, written from scratch, bound to be teething problems, but its looking as though it will be fine long run when the CPU utilisation is resolved. Its a pity the app was not run as a separate choice in preferences as a Public Beta, as at present being Prime Time app there's a lot of pressure to get it to "perform".

It'll happen, meanwhile for those not wanting to run a CPU WU, all appears fine.

Regards
Zy

Thanks ZY, you've confirmed what I had more or less surmised for myself, that CPU projects are no go at the moment on MW GPU rigs. Oh well, I'll just have to put up with it for the time being I suppose. I hope the credit we are going to be getting is gonna reflect that though :-))

for a majority of folks so far, this certainly seems to be the general consensus. however, keep in mind that the ability of a rig to crunch CPU tasks and MW@H GPU tasks simultaneously (and do so without devoting excessive CPU resources to the GPU and crippling GPU task run times) depends on several variables...hence the reason so many different platforms are acting so differently. for instance, as i mentioned several posts above, i experience a ~20% decrease in MW@H task efficiency while running Einstein@Home on all 6 CPU cores. and yet i do not need to suspend E@H altogether in order to achieve "normal" MW@H task run times. rather i simply have to tell BOINC to only use 85% of the CPU, leaving one CPU core free to handle the uploading and offloading of MW@H data to and from the GPU (and to handle any other multitasking instructions that may occur while i browse the web or whatever). granted, this is made possible by the fact that my "CPU utilization per GPU task" is the normal 5% (0.05 CPUs + 1.00 ATI GPUs) and not excessive like so many others are experiencing.

so for those of you who are experiencing significant slow-downs in MW@H task run times due to BOINC CPU tasks using up all of your CPU cores (100% CPU load), i would try disabling a core at a time in BOINC before i suspend an entire project that crunches CPU tasks. that is, see if you can at least load some of your CPU with BOINC CPU tasks from other projects without hindering MW@H GPU performance.

i understand that this is just a temporary fix to allow those who are capable to run CPU tasks and MW@H GPU tasks simultaneously, and that ultimately the current MW@H app will have to be further revised so that we can utilize all of our CPU cores on other projects without taking a MW@H GPU task performance hit. but i thought it was worth mentioning for those trying to minimize their slow-downs and losses...

I am running two workunits and it does increase the GPU usage up to 100% and the CPU usage is still at 0%

How are you running two workunits concurrently ?

perhaps he has 2 GPU clients in the same rig

---