From 10e9a53864e1f49746c605b4beb76af96fc8e6fe Mon Sep 17 00:00:00 2001 From: eliottrosenberg <61400172+eliottrosenberg@users.noreply.github.com> Date: Sun, 11 Feb 2024 02:31:04 -0500 Subject: [PATCH 1/4] Add option to use image with CUDA preinstalled --- docs/tutorials/gcp_gpu.md | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/docs/tutorials/gcp_gpu.md b/docs/tutorials/gcp_gpu.md index 73d26b97..a072b296 100644 --- a/docs/tutorials/gcp_gpu.md +++ b/docs/tutorials/gcp_gpu.md @@ -35,6 +35,11 @@ instance section, ensure that your VM has the following properties: 1. In the **Operating System** option, choose **Ubuntu**. 2. In the **Version** option, choose **20.04 LTS**. 3. In the **Size** field, enter **40** (minimum). + + **Alternatively, you can click the "Switch Image" button and use the image with + CUDA pre-installed, which lets you skip step 3. I have verified that this works + with cuQuantum Appliance (option 3).** + * The instructions above override steps 3 through 5 in the [Create a Linux VM instance](https://cloud.google.com/compute/docs/quickstart-linux) Quickstart. @@ -176,8 +181,8 @@ to set up NVIDIA Container Toolkit. ## 9. Install NVIDIA cuQuantum Appliance (Option 3 only) Follow the instructions [here](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/cuquantum-appliance) -to set up cuQuantum Appliance. You may need to use `sudo` for the Docker commands. - +to set up cuQuantum Appliance. If this link is broken, you can try the [cached page](https://webcache.googleusercontent.com/search?q=cache%3Ahttps%3A%2F%2Fcatalog.ngc.nvidia.com%2Forgs%2Fnvidia%2Fcontainers%2Fcuquantum-appliance&rlz=1CAFOSO_enUS1067&oq=cache%3Ahttps%3A%2F%2Fcatalog.ngc.nvidia.com%2Forgs%2Fnvidia%2Fcontainers%2Fcuquantum-appliance&gs_lcrp=EgZjaHJvbWUqBggAEEUYOzIGCAAQRRg7MgYIARBFGDoyBggCEEUYPDIGCAMQRRg80gEIMzMyMGowajSoAgCwAgA). +You may need to use `sudo` for the Docker commands. ## 10. Verify your installation (Options 1, 2, and 3) From 510c6587833d582f9af36c3c27089780c1409e56 Mon Sep 17 00:00:00 2001 From: eliottrosenberg <61400172+eliottrosenberg@users.noreply.github.com> Date: Sun, 11 Feb 2024 22:08:58 -0500 Subject: [PATCH 2/4] Update number of V100s It seems that 8 V100s cannot share memory, but 4 can. --- docs/choose_hw.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/choose_hw.md b/docs/choose_hw.md index 8f931082..c57159a2 100644 --- a/docs/choose_hw.md +++ b/docs/choose_hw.md @@ -140,8 +140,8 @@ At the time of writing, the following compatible GPUs are available on GCP: * [NVIDIA V100](https://www.techpowerup.com/gpu-specs/tesla-v100-pcie-16-gb.c2957). Like the NVIDIA T4, this GPU has 16GB of RAM and therefore supports up to 30 qubits. It is faster than the T4. - Further, it is compatible with multi-GPU simulations. With 8 NVIDIA V100s (128GB), - you can simulate up to 33 qubits. + Further, it is compatible with multi-GPU simulations. With 4 NVIDIA V100s (64GB), + you can simulate up to 32 qubits. * [NVIDIA L4](https://www.techpowerup.com/gpu-specs/l4.c4091). This GPU has 24GB of RAM and can therefore simulate up to 31 qubits. With eight of them (192GB), you can simulate up to 34 qubits. From 38a09094f6a985731efcaeab6600f7c835e825d0 Mon Sep 17 00:00:00 2001 From: eliottrosenberg <61400172+eliottrosenberg@users.noreply.github.com> Date: Mon, 12 Feb 2024 10:43:12 -0500 Subject: [PATCH 3/4] implement suggestions --- docs/tutorials/gcp_gpu.md | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/docs/tutorials/gcp_gpu.md b/docs/tutorials/gcp_gpu.md index a072b296..5669e39c 100644 --- a/docs/tutorials/gcp_gpu.md +++ b/docs/tutorials/gcp_gpu.md @@ -37,7 +37,7 @@ instance section, ensure that your VM has the following properties: 3. In the **Size** field, enter **40** (minimum). **Alternatively, you can click the "Switch Image" button and use the image with - CUDA pre-installed, which lets you skip step 3. I have verified that this works + CUDA pre-installed, which lets you skip step 3. It has been verified that this works with cuQuantum Appliance (option 3).** * The instructions above override steps 3 through 5 in the [Create a Linux VM @@ -181,8 +181,7 @@ to set up NVIDIA Container Toolkit. ## 9. Install NVIDIA cuQuantum Appliance (Option 3 only) Follow the instructions [here](https://catalog.ngc.nvidia.com/orgs/nvidia/containers/cuquantum-appliance) -to set up cuQuantum Appliance. If this link is broken, you can try the [cached page](https://webcache.googleusercontent.com/search?q=cache%3Ahttps%3A%2F%2Fcatalog.ngc.nvidia.com%2Forgs%2Fnvidia%2Fcontainers%2Fcuquantum-appliance&rlz=1CAFOSO_enUS1067&oq=cache%3Ahttps%3A%2F%2Fcatalog.ngc.nvidia.com%2Forgs%2Fnvidia%2Fcontainers%2Fcuquantum-appliance&gs_lcrp=EgZjaHJvbWUqBggAEEUYOzIGCAAQRRg7MgYIARBFGDoyBggCEEUYPDIGCAMQRRg80gEIMzMyMGowajSoAgCwAgA). -You may need to use `sudo` for the Docker commands. +to set up cuQuantum Appliance. You may need to use `sudo` for the Docker commands. ## 10. Verify your installation (Options 1, 2, and 3) From 48aea853308ab8b523beb4b853a81cb68d318954 Mon Sep 17 00:00:00 2001 From: eliottrosenberg <61400172+eliottrosenberg@users.noreply.github.com> Date: Mon, 12 Feb 2024 11:17:37 -0500 Subject: [PATCH 4/4] update link --- docs/choose_hw.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/choose_hw.md b/docs/choose_hw.md index c57159a2..19094b04 100644 --- a/docs/choose_hw.md +++ b/docs/choose_hw.md @@ -129,7 +129,7 @@ choose a specific machine: For GPU simulations, you may follow the instructions in [this](tutorials/gcp_gpu) guide to set up a virtual machine (VM) on Google Cloud Platform (GCP). Alternatively, you can use your own hardware. -Note the [hardware requirements](https://docs.nvidia.com/cuda/cuquantum/getting_started.html#custatevec) +Note the [hardware requirements](https://docs.nvidia.com/cuda/cuquantum/latest/getting_started.html#custatevec) for NVIDIA's cuQuantum when picking a GPU; in particular, it must have CUDA Compute Capability 7.0 or higher. At the time of writing, the following compatible GPUs are available on GCP: