How Can I change device order in multiGPU?

Accelerated Computing CUDA CUDA Programming and Performance
1

I’ve installed 4 GPUs.

I installed 3 C870 for computation and 1 8600GT for display.
and it works well but, the device 0 is 8600GT and 1,2,3 is C870…
I want to device 0,1,2 for C870 and 3 for 8600GT.

My linux is fedora 8 and installed 169.09 driver.
What to do? thanks.

====================================

$ ./deviceQuery
There are 4 devices supporting CUDA

Device 0: “GeForce 8600 GT”
Major revision number: 1
Minor revision number: 1
Total amount of global memory: 1073020928 bytes
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 8192
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 262144 bytes
Texture alignment: 256 bytes
Clock rate: 1188000 kilohertz

Device 1: “Tesla C870”
Major revision number: 1
Minor revision number: 0
Total amount of global memory: 1610350592 bytes
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 8192
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 262144 bytes
Texture alignment: 256 bytes
Clock rate: 1350000 kilohertz

Device 2: “Tesla C870”
Major revision number: 1
Minor revision number: 0
Total amount of global memory: 1610350592 bytes
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 8192
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 262144 bytes
Texture alignment: 256 bytes
Clock rate: 1350000 kilohertz

Device 3: “Tesla C870”
Major revision number: 1
Minor revision number: 0
Total amount of global memory: 1610350592 bytes
Total amount of constant memory: 65536 bytes
Total amount of shared memory per block: 16384 bytes
Total number of registers available per block: 8192
Warp size: 32
Maximum number of threads per block: 512
Maximum sizes of each dimension of a block: 512 x 512 x 64
Maximum sizes of each dimension of a grid: 65535 x 65535 x 1
Maximum memory pitch: 262144 bytes
Texture alignment: 256 bytes
Clock rate: 1350000 kilohertz

Test PASSED

Press ENTER to exit…

2

You could try changing the order in which the cards are in the PCIe slots, but that may not even work. This ordering is completely determined in the driver.

3

I’ve tried this method, but I’ve failed

some days ago, I’ve got device 0 to 3 S870 and device 4 for 8800gtx with ordered PCIe slots in another system. but, I don’t remember when I does work.

I’ll try this…

step0 : In this situation, 8600gt widh device 0 and 3 C870 with device 1,2,3

step1 : I’ll install 3 C870 with ATI graphic cards then, C870 will be 0,1,2

( with which device for X-windows? I don’t remember …)

step2 : I’ll replace from ATI to 8600GT

step3: deviceQuery

I guess it was my setting in S870 and 8800GTX in the machine. I’ll test it later.

4

I’ve captured deviceQuery results with s870 ( 4 C870) & ATI graphic card

Device 0: “Tesla C870”

Major revision number: 1

Minor revision number: 0

Total amount of global memory: 1610350592 bytes

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 16384 bytes

Total number of registers available per block: 8192

Warp size: 32

Maximum number of threads per block: 512

Maximum sizes of each dimension of a block: 512 x 512 x 64

Maximum sizes of each dimension of a grid: 65535 x 65535 x 1

Maximum memory pitch: 262144 bytes

Texture alignment: 256 bytes

Clock rate: 1350000 kilohertz

Device 1: “Tesla C870”

Major revision number: 1

Minor revision number: 0

Total amount of global memory: 1610350592 bytes

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 16384 bytes

Total number of registers available per block: 8192

Warp size: 32

Maximum number of threads per block: 512

Maximum sizes of each dimension of a block: 512 x 512 x 64

Maximum sizes of each dimension of a grid: 65535 x 65535 x 1

Maximum memory pitch: 262144 bytes

Texture alignment: 256 bytes

Clock rate: 1350000 kilohertz

Device 2: “Tesla C870”

Major revision number: 1

Minor revision number: 0

Total amount of global memory: 1610350592 bytes

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 16384 bytes

Total number of registers available per block: 8192

Warp size: 32

Maximum number of threads per block: 512

Maximum sizes of each dimension of a block: 512 x 512 x 64

Maximum sizes of each dimension of a grid: 65535 x 65535 x 1

Maximum memory pitch: 262144 bytes

Texture alignment: 256 bytes

Clock rate: 1350000 kilohertz

Device 3: “Tesla C870”

Major revision number: 1

Minor revision number: 0

Total amount of global memory: 1610350592 bytes

Total amount of constant memory: 65536 bytes

Total amount of shared memory per block: 16384 bytes

Total number of registers available per block: 8192

Warp size: 32

Maximum number of threads per block: 512

Maximum sizes of each dimension of a block: 512 x 512 x 64

Maximum sizes of each dimension of a grid: 65535 x 65535 x 1

Maximum memory pitch: 262144 bytes

Texture alignment: 256 bytes

Clock rate: 1350000 kilohertz

Test PASSED

Press ENTER to exit…