This host (AMD 1090t, Win7-64, HD 5870, BOINC 6.12.34) is back in business after upgrading to Cat 12.1. I don't know if any of the 11.x versions would have worked -- I went straight up from 10.10. No other changes were made. WU's now complete 4-5% faster (2-3 sec) on average.

---

- The old CAL version is gone; it's replaced with the OpenCL application

Question: Has the nVidia application also switched from CUDA to OpenCL now? The application name suggests it.

Seems to work fine on a GTX 560 Ti w/ Win7 64bit and BOINC 6.12.34 (x64), first result is valid. The ATI onboard GPU (SP) is not causing problems, server tells me: "An ATI GPU supporting double precision math is required" and "ATI GPU R600 (R38xx) does not support OpenCL".


<core_client_version>6.12.34</core_client_version>
<![CDATA[
<stderr_txt>
<search_application> milkyway_separation 1.02 Windows x86_64 double OpenCL </search_application>
Guessing preferred OpenCL vendor 'NVIDIA Corporation'
Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
Using SSE3 path
Found 1 platform
Platform 0 information:
Name: NVIDIA CUDA
Version: OpenCL 1.1 CUDA 4.1.1
Vendor: NVIDIA Corporation
Extensions: cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_d3d9_sharing cl_nv_d3d10_sharing cl_khr_d3d10_sharing cl_nv_d3d11_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll
Profile: FULL_PROFILE
Using device 0 on platform 0
Found 1 CL device
Device 'GeForce GTX 560 Ti' (NVIDIA Corporation:0x10de) (CL_DEVICE_TYPE_GPU)
Driver version: 285.62
Version: OpenCL 1.1 CUDA
Compute capability: 2.1
Max compute units: 8
Clock frequency: 1645 Mhz
Global mem size: 2147483648
Local mem size: 49152
Max const buf size: 65536
Double extension: cl_khr_fp64
Build log:
--------------------------------------------------------------------------------

ptxas info : Compiling entry function 'probabilities' for 'sm_21'
ptxas info : Function properties for probabilities
80 bytes stack frame, 76 bytes spill stores, 76 bytes spill loads
ptxas info : Used 62 registers, 100 bytes cmem[0], 56 bytes cmem[16]
--------------------------------------------------------------------------------
Build log:
--------------------------------------------------------------------------------

ptxas info : Compiling entry function 'summarization' for 'sm_21'
ptxas info : Function properties for summarization
0 bytes stack frame, 0 bytes spill stores, 0 bytes spill loads
ptxas info : Used 13 registers, 2048+0 bytes smem, 48 bytes cmem[0]
--------------------------------------------------------------------------------
Estimated Nvidia GPU GFLOP/s: 842 SP GFLOP/s, 105 DP FLOP/s
Using a target frequency of 60.0
Using a block size of 4096 with 17 blocks/chunk
Using clWaitForEvents() for polling with initial wait of 12 ms (mode 0)
Range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Iteration area: 2240000
Chunk estimate: 32
Num chunks: 33
Chunk size: 69632
Added area: 57856
Effective area: 2297856
Initial wait: 12 ms
Integration time: 542.787896 s. Average time per iteration = 848.106088 ms
Integral 0 time = 546.308173 s
Running likelihood with 109999 stars
Likelihood time = 2.219000 s
<background_integral> 0.000133824638386 </background_integral>
<stream_integral> 6.443820669765729 875.872833232415470 </stream_integral>
<background_likelihood> -2.940328087823295 </background_likelihood>
<stream_only_likelihood> -83.771552699247522 -7.323208286845183 </stream_only_likelihood>
<search_likelihood> -2.935410274019145 </search_likelihood>
11:38:38 (3120): called boinc_finish

</stderr_txt>
]]>

Thanks Arkayn and Maac59 for your suggestions and link.
It is working smoothly again under Win7-64ultimate, BOINC 6.10.58 and CAL 12.1.

---

Greetings from,
TJ

Could be -- I've over a dozen 4850 systems in play -- then again, it might encourage me to simply shut down a batch of systems and reduce the electric bill. I tend to do that in the summer anyway.

Prices are bound to come down, so dont be too depressed, there's no way they will stay this high for good. Also another new series like 8XXX are probably 12-18 months away, that will also help depress prices of 5XXX and 6XXX both of which have full OpenCL support.

Personal guess is it becomes a real drama circa 18months from now when all have gone OpenCL (as such), but could be that future series will knock down 5XXX and 6XXX into more affordable scenarios by then. Difficult to say, but on balance I reckon it will trend that way with second prices of 5XXX and 6XXX.

Regards
Zy

Barry maybe EBay will become your friend and help you unload those cards and get newer, faster ones! Just remember EVERYONE on EBay is NOT honest!!!

This host (AMD 1090t, Win7-64, HD 5870, BOINC 6.12.34) is back in business after upgrading to Cat 12.1. I don't know if any of the 11.x versions would have worked -- I went straight up from 10.10. No other changes were made. WU's now complete 4-5% faster (2-3 sec) on average.

I am using 11.9 on all of my pc's and I am getting work on two of them! I only have 5870's on my gpu's here at MW, so that is the same but the Boinc version is different and so is whether they are 32 or 64 bit. I re-enabled work so I will wait and see if it picks some up later. I also did a reset on one but nothing was happening, so I left it alone and turned off the monitor but the pc is still running.

I found that once I had everything set to go, I needed to just leave it alone. The WUs came in one at a time and it took a few minutes to get the next one. I had to force myself not to babysit and manually update Boinc trying to get more work every time it ran out. I let it go overnight with no other GPU projects running. By morning it was running just as before.
Win 7-64
Boinc 7.0.8(x64)
Catalyst 11.9
HD5870 GPU speed = 900MHZ - memory speed = 300MHz
OpenCL app from Arkayn
With an app_info.xml instruction to run two WUs at a time they complete in ~2:12 minutes with approx 5 sec CPU time.

The Nvidia app has always been OpenCL. We used to be a ATI only project.

---

This host (AMD 1090t, Win7-64, HD 5870, BOINC 6.12.34) is back in business after upgrading to Cat 12.1...

For completeness, I might maybe should have added that I'm using the stock 1.02 from the project, no app_info file.

---

For completeness, I might maybe should have added that I'm using the stock 1.02 from the project, no app_info file.

the app_info.xml file really isn't necessary unless you either want to change the number of simultaneously running MW@H tasks to something greater than 1 or improve GUI lag/set process priority/etc. via command line parameters.

---

I said 2 days ago:

As a point of information my HD 4770 (11.6, BOINC 6.12.34, winXP32, AMD64 3800+) has been very unhappy with the change to OpenCL. Host 179083 ( http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=179083 ) gets OpenCL work, but promptly errors out with -1073740777 (unknown error). I detached and tried again with the same result.

Upgrading from Cat 11.6 to Cat 12.1 made MW work without trashing anything. Of course Matt has made a number of changes in the meantime. Moo works too. I wasn't sure Cat 12.1 would support an (old) HD 4770 adequately. I guess it does.

Performance is very near what it was with the CAL app. I've seen a decrease of about 4 seconds from 192 seconds total time, but an increase of 3 seconds (from 6 to 9 seconds) of CPU time. I'd have never noticed if I wasn't looking.

An appropriate classic quote of wisdom:

"Thank god Mr. Spock, you've fixed it"
"There was no deity involved Mr. Scott. It was my cross circuiting from A to B that did it"

For completeness, I might maybe should have added that I'm using the stock 1.02 from the project, no app_info file.

the app_info.xml file really isn't necessary unless you either want to change the number of simultaneously running MW@H tasks to something greater than 1 or improve GUI lag/set process priority/etc. via command line parameters.

What command would be added to the the app-info file to throttle back GPU usage? It's currently at 99% and I'd like to pull it back to 80-85%.
Win7 64 bit. Radeon 5830.

---

For completeness, I might maybe should have added that I'm using the stock 1.02 from the project, no app_info file.

the app_info.xml file really isn't necessary unless you either want to change the number of simultaneously running MW@H tasks to something greater than 1 or improve GUI lag/set process priority/etc. via command line parameters.

What command would be added to the the app-info file to throttle back GPU usage? It's currently at 99% and I'd like to pull it back to 80-85%.
Win7 64 bit. Radeon 5830.

sorry...i've only done it successfully on WinXP 32-bit like you. i have yet to tackle the issue on Win7 x64, let alone install it on one of my machines...but hopefully i'll be getting a licensed copy from work soon...

---

What command would be added to the the app-info file to throttle back GPU usage? It's currently at 99% and I'd like to pull it back to 80-85%.
Win7 64 bit. Radeon 5830.

With the new app you don't need the app_info-file to control GPU load - it's now only necessary if you want to run multiple instances on one GPU. If you want to use it anyway the command line parameter apparently wasn't changed from the old app:
--gpu-target-frequency XX

Raising XX lessens the GPU load and increases the run time (default is now 60).

With the new app it's easier controlled in the Milkyway@home preferences: Frequency (in Hz) that should try to complete individual work chunks

---

Starfire

Thanks

---

Thanks, how do you think I was able to afford the dozen or so 4850's -- I well know about ebay -- and look closely at listings there.

Most of the 4850's I got were fine, but a couple of them needed replacement heatsink and fan combinations -- which I keep around just in case.

I also got one 4870 from NewEgg -- but for whatever reason, that puppy runs GPU BOINC applications VERY slow. I just swapped out a re-fanned 4850 there.

In any event, after my summer system shutdowns are over this coming fall, I may look into getting the next series up. What 5xxx AMD cards run double precision -- I know their numbering scheme got a bit different (ie a 5850 might well not be double precision).

Barry maybe EBay will become your friend and help you unload those cards and get newer, faster ones! Just remember EVERYONE on EBay is NOT honest!!!

I've got an AMD FX-8150 CPU, is there anything I need to modify to receive the new WU's?

- AVX will be used if available on Linux

I'm using BOINC 7.0.15 in OpenSUSE 12.1

5850 is double precision.

---

In any event, after my summer system shutdowns are over this coming fall, I may look into getting the next series up. What 5xxx AMD cards run double precision -- I know their numbering scheme got a bit different (ie a 5850 might well not be double precision).

The 58xx series does have double precision, it is the 68xx series that does not have it.

We are still waiting to see what happens with the 78xx series.

---

My ATI 4850 is not processing the new work units. In an attempt to make the suggested upgrade, my 4850 is now not even processing the Collatz projects that it was processing fine. My machine is running an ATI 3300, 4850 and Nvidia 550. I am afraid I am returning more errors now than work.

Mark, the combination I've found that works for me with 4850's.

1) AMD video1 1.8
2) BOINC 6.10.58

This works with both Collatz and MW -- as well as MooWrapper.

I've used this with XP, Win7 32 and Win7 64 bit setups.

Also, after the software versions are in shape, I'd detatch and reattach the projects.

My ATI 4850 is not processing the new work units. In an attempt to make the suggested upgrade, my 4850 is now not even processing the Collatz projects that it was processing fine. My machine is running an ATI 3300, 4850 and Nvidia 550. I am afraid I am returning more errors now than work.