What to fold with only one of two GPUs
Posted: Sun Dec 13, 2015 2:44 am
Hello fellow Folding@home users ,
This is my first post on the Folding@home support forum. I'm still relatively new (less than one year) to Folding@home. I have two GPUs and I want Folding@home to just use one of them. Under Configure > Slots > I have removed the additional GPU and the CPU, I only have one GPU remaining. I have edited this so that the GPU-index = 1, opencl-index = 1 and cuda-index = 1. I have pressed okay, saved the changes and restarted the computer only to find that the GPU-index has reverted to -1 and it is once again using the Geforce GT 610.
GPU0 (Folding@home) GPU0 (GPU-Z) GPU0 (FAHClient.exe –lspci) PCIe x16_1 (physical motherboard location) Geforce GT 610
GPU1 (Folding@home) GPU1 (GPU-Z) GPU1 (FAHClient.exe –lspci) PCIe x16_2 (physical motherboard location) Geforce GTX 750 Ti
I was sure I was correct by declaring the following under the slot option: GPU-index = 1, opencl-index = 1 and cuda-index = 1. It seems to recognise it when I do this in the client on the face of it. But judging from the temperatures in TThrottle for GPU0 and GPU1 (and the very slow progress) it is actually stubbornly using GPU0 (Geforce GT 610) rather than GPU1 (Geforce GTX 750 Ti). When I reboot, as mentioned above the client just goes to GPU-index = -1, opencl-index = 1 and cuda-index = 1. The client acknowledges only the GPU0 Geforce GT 610 when it comes to processing work units and recommences (usually with a new work unit ) the very slow folding. I'd really welcome any assistance anyone may be able to offer to help me resolve this problem.
Chris
Further details about my system and situation...
SLI is disabled.
GPU-Z has ticks for OpenCL and CUDA for both GPUs.
Computer Specs:
CPU: Intel i7 CPU 950@3.07GHz
4 cores 8 threads
Memory: 12GB DDR3
Motherboard: ASUS P6T
Power Supply: 550W
OS: Windows 7
GPU0: Geforce GT 610
GPU1: Geforce GTX 750 Ti
Network Connection:
ADSL
Wireless network adapter in PCIe x16_3 slot
Overclocked?:
There is some automated overclocking controlled by the motherboard up to 3.2GHz or so. The CPU has a closed loop water cooler. There are plenty of fans in the case too (I can't fit any more in).
Stable?:
Without Folding@home running my system seems stable. It's still a bit of an open question as to whether or not the system is stable running as I would like it to be. The PC is stable overall.
Software:
I am using version 7.4.4 of the Folding@home client. I also have BOINC running and configured to use the Geforce GT 610 and some of the threads (87.5% of the CPU resources, 7 threads), leaving one for the GPU I hope to use with Folding@home. I am not running BOINC for the moment though, while I try to get Folding@home to use the right GPU.
WU details:
I'm not sure what to put here. I'm hoping Folding@home will recognise my desired configuration for this computer and send me the appropriate work units. Preferably something I can complete in a 12 hour interval or something that will allow me to save the progress I've made. While I would like to leave my computer running longer, I can't.
This is my first post on the Folding@home support forum. I'm still relatively new (less than one year) to Folding@home. I have two GPUs and I want Folding@home to just use one of them. Under Configure > Slots > I have removed the additional GPU and the CPU, I only have one GPU remaining. I have edited this so that the GPU-index = 1, opencl-index = 1 and cuda-index = 1. I have pressed okay, saved the changes and restarted the computer only to find that the GPU-index has reverted to -1 and it is once again using the Geforce GT 610.
GPU0 (Folding@home) GPU0 (GPU-Z) GPU0 (FAHClient.exe –lspci) PCIe x16_1 (physical motherboard location) Geforce GT 610
GPU1 (Folding@home) GPU1 (GPU-Z) GPU1 (FAHClient.exe –lspci) PCIe x16_2 (physical motherboard location) Geforce GTX 750 Ti
I was sure I was correct by declaring the following under the slot option: GPU-index = 1, opencl-index = 1 and cuda-index = 1. It seems to recognise it when I do this in the client on the face of it. But judging from the temperatures in TThrottle for GPU0 and GPU1 (and the very slow progress) it is actually stubbornly using GPU0 (Geforce GT 610) rather than GPU1 (Geforce GTX 750 Ti). When I reboot, as mentioned above the client just goes to GPU-index = -1, opencl-index = 1 and cuda-index = 1. The client acknowledges only the GPU0 Geforce GT 610 when it comes to processing work units and recommences (usually with a new work unit ) the very slow folding. I'd really welcome any assistance anyone may be able to offer to help me resolve this problem.
Chris
Further details about my system and situation...
SLI is disabled.
GPU-Z has ticks for OpenCL and CUDA for both GPUs.
Computer Specs:
CPU: Intel i7 CPU 950@3.07GHz
4 cores 8 threads
Memory: 12GB DDR3
Motherboard: ASUS P6T
Power Supply: 550W
OS: Windows 7
GPU0: Geforce GT 610
GPU1: Geforce GTX 750 Ti
Network Connection:
ADSL
Wireless network adapter in PCIe x16_3 slot
Overclocked?:
There is some automated overclocking controlled by the motherboard up to 3.2GHz or so. The CPU has a closed loop water cooler. There are plenty of fans in the case too (I can't fit any more in).
Stable?:
Without Folding@home running my system seems stable. It's still a bit of an open question as to whether or not the system is stable running as I would like it to be. The PC is stable overall.
Software:
I am using version 7.4.4 of the Folding@home client. I also have BOINC running and configured to use the Geforce GT 610 and some of the threads (87.5% of the CPU resources, 7 threads), leaving one for the GPU I hope to use with Folding@home. I am not running BOINC for the moment though, while I try to get Folding@home to use the right GPU.
WU details:
I'm not sure what to put here. I'm hoping Folding@home will recognise my desired configuration for this computer and send me the appropriate work units. Preferably something I can complete in a 12 hour interval or something that will allow me to save the progress I've made. While I would like to leave my computer running longer, I can't.