I'm testing a new framework for running our searches, which should make them a bit more reliable and reduce server load. I've sent out a small batch of 1000 (and won't be generating new ones) to see if this works. Let me know if they crunch correctly for you, they should be named 'ps_test_1_X' where X is some number.

---

All my test units ps_test_1_ have computational errors

Regards

---

I only use my GPU on this project but all computations for that fail immediately.

Regards

edit: I mean the test units ofc, rest is fine ;)

All ps_test_1_X WUs failed immediately computation error with my ATI 4850

yeah, these are borked. ps_test_2 should work.

---

If I understand what are units GPU? améliroration one of the current application?

for now I only get the de_separation ....

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

OK, currently working on regular WU's. Once I'll get the test_2 units ill let you know.

Or... maybe ps_test_3

---

If I understand what are units GPU? améliroration one of the current application?

They're separation workunits, so they should be CPU/GPU workunits. They aren't for the n-body application.

---

All erroring out here. NNT for now.

Since I won't become UOTD anyway for my charming posts: Wasn't there a roadmap saying a) get things stable b) tackle the lag if possible c) do the credit x2 thingy?

Hmm.

If I understand what are units GPU? améliroration one of the current application?

They're separation workunits, so they should be CPU/GPU workunits. They aren't for the n-body application.

????

These are units GPU or CPU?

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

If I understand what are units GPU? améliroration one of the current application?

They're separation workunits, so they should be CPU/GPU workunits. They aren't for the n-body application.

????

These are units GPU or CPU?

Both CPU and GPU. The 'separation' workunits can run on either. The n-body workunits can only run on the CPU.

---

All erroring out here. NNT for now.

Since I won't become UOTD anyway for my charming posts: Wasn't there a roadmap saying a) get things stable b) tackle the lag if possible c) do the credit x2 thingy?

Hmm.

I'm only putting them out in small batches and checking to see if they work. No need to go NNT, as all the 'de_separation_' workunit will still being generated and sent out.

Either way, looks like the 'ps_test_4_X' workunits are crunching and returning results.

---

ok for them, using CPU and GPU ....

for now I'm only units and n_body de_separation usual ....

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

ok for them, using CPU and GPU ....

for now I'm only units and n_body de_separation usual ....

That's fine, I'm not sending out very many of these, just a few for testing to make sure they work.

---

So in preferences should check any CPU and GPU?

Edit : Oupss I had forgotten to remove the file app_info .... to receive these units ...

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

Test4 seems to be working fine to completion

Regards

Received 2 test_4 Wu's and both crunched to completion on my GPU.

Regards

for now, one is marked as finished and complete marked invalid ....

<core_client_version> 6.12.27 </ core_client_version>
<! [CDATA [
<stderr_txt>
<search_application> milkywayathome separation 0.52 Windows x86 dual OpenCL </ search_application>
Found 1 platforms
0 Information Platform:
Platform name: NVIDIA CUDA
Platform Version: 4.0.1 OpenCL CUDA 1.0
Platform vendor:
Platform profile:
Platform 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
Using one platform device 0 0
Found 1 LC devices
Device GeForce GTX 470 (NVIDIA Corporation: 0x10de)
Type: CL_DEVICE_TYPE_GPU
Driver version: 275.27
Version: CUDA 1.0 OpenCL
Compute capability: 2.0
Little endian: CL_TRUE
Error correction: CL_FALSE
Image support: CL_TRUE
Address bits: 32
Max Computer unit: 14
Clock frequency: 1312 MHz
Global Mem size: 1275658240
Max mem alloc: 318914560
Global mem cache: 229376
Cacheline size: 128
Local Member type: CL_LOCAL
Local mem size: 49152
Const max args: 9
Const max buf size: 65536
Parameter max size: 4352
Max work size group: 1024
Max work item Sun: 3
Max work item sizes: {1024, 1024, 64}
Mem base addr align: 4096
Min align type size: 128
Timer resolution: 1000 ns
Double extension: MW_CL_KHR_FP64
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 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64

Compile flags:
-Cl-cl-enable-mad-no-signed-zeros-cl-cl-strict-aliasing-finite-math-only-0-DUSE_MAD DUSE_CL_MATH_TYPES = = = DUSE_FMA 1-0-cl-nv-verbose-DDOUBLEPREC = 1 =-DMILKYWAY_MATH_COMPILATION DNSTREAM-3-1-DAUX_BG_PROFILE DFAST_H_PROB = = = 0-1-DUSE_IMAGES DI_DONT_KNOW_WHY_THIS_DOESNT_WORK_HERE = 1

Build status: CL_BUILD_SUCCESS
Build log:


Kernel Work Group Info:
Work group size = 512
Kernel local mem size = 0
Compiles work group size = {0, 0, 0}
Group size = 64, per UC = 32, UC = 2048 threads per
Block size = 28672
Desired = 40
Min Ground: 40 53760
Lower solution n: n = 40, x = 53760
Higher solution n: n = 40, x = 53760
Using solution: n = 40, x = 53760
Range: {nu_steps = 640, mu_steps = 1600, r_steps = 1400}
Iteration area: 2240000
Chunk estimate: 40
Num chunks: 40
Added area: 53,760
Effective area: 2293760
Block size: 28672
Integration time: 428.279406 seconds Average time per iteration = 669.186571 ms
<background_integral> 0.00054795287962228472 </ background_integral>
<stream_integrals> 226.40062096088954000000 168.70556324250637000000 2029.83613571773840000000 </ stream_integrals>
<background_only_likelihood> -3.65102064360549640000 </ background_only_likelihood>
<stream_only_likelihood> -6.03628400706629440000 -4.14500894539928130000 -4.28000110367938550000 </ stream_only_likelihood>
<search_likelihood> -3.13443262293875690000 </ search_likelihood>
3:17:51 p.m. (2592): Called boinc_finish

</ Stderr_txt>
]]>

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

Hey! switch that machina back on! I completed my 12 WU,s and my GPU is cooling down!