From a8f004043fd9084f529f88d2ce7798536014036a Mon Sep 17 00:00:00 2001 From: surcyf123 Date: Sun, 20 Sep 2026 04:30:54 +0000 Subject: [PATCH 1/2] Add Lium (lium.io) --- docs/cloud-gpus/cloud-gpus.csv | 15 ++++++++++++++- docs/cloud-gpus/index.md | 2 +- 2 files changed, 15 insertions(+), 2 deletions(-) diff --git a/docs/cloud-gpus/cloud-gpus.csv b/docs/cloud-gpus/cloud-gpus.csv index a92ed357..57794d79 100644 --- a/docs/cloud-gpus/cloud-gpus.csv +++ b/docs/cloud-gpus/cloud-gpus.csv @@ -280,4 +280,17 @@ Thunder Compute,A100 (80 GB),Ampere,8,640,94,648,11.92,1.49,, Thunder Compute,H100 (80 GB),Hopper,1,80,4,32,1.38,2.19,, Thunder Compute,H100 (80 GB),Hopper,2,160,8,64,4.38,2.19,, Thunder Compute,H100 (80 GB),Hopper,4,320,60,480,11.56,2.89,, -Thunder Compute,H100 (80 GB),Hopper,8,640,120,960,23.12,2.89,, \ No newline at end of file +Thunder Compute,H100 (80 GB),Hopper,8,640,120,960,23.12,2.89,, +Lium,B300 (288 GB),Blackwell,1,288,24,340,8.00,8.00,, +Lium,B300 (288 GB),Blackwell,8,2304,256,3023,64.00,8.00,, +Lium,B200 (180 GB),Blackwell,1,180,26,354,5.60,5.60,, +Lium,H200 (141 GB),Hopper,1,141,24,236,3.00,3.00,, +Lium,H200 (141 GB),Hopper,4,564,128,1008,13.60,3.40,, +Lium,H200 (141 GB),Hopper,8,1128,128,1344,31.92,3.99,, +Lium,H100 (80 GB),Hopper,1,80,20,236,1.30,1.30,, +Lium,H100 (80 GB),Hopper,2,160,384,503,2.82,1.41,, +Lium,H100 (80 GB),Hopper,8,640,208,1772,22.00,2.75,, +Lium,RTX PRO 6000 (96 GB),Blackwell,1,96,16,142,1.19,1.19,, +Lium,RTX PRO 6000 (96 GB),Blackwell,8,768,256,1007,14.32,1.79,, +Lium,RTX 5090 (32 GB),Blackwell,1,32,16,30,0.58,0.58,, +Lium,RTX 5090 (32 GB),Blackwell,8,256,384,377,4.00,0.50,, \ No newline at end of file diff --git a/docs/cloud-gpus/index.md b/docs/cloud-gpus/index.md index 7964d9c5..083b7749 100644 --- a/docs/cloud-gpus/index.md +++ b/docs/cloud-gpus/index.md @@ -37,7 +37,7 @@ The table below does not include all possible configurations for all providers, - Most providers, including AWS, Azure, and Lambda, provide instances with pre-set configurations. - On GCP, any suitable machine can be connected to a configuration of GPUs. -- On other providers, like Oblivus Cloud, Cudo Compute, and RunPod, users have precise control over the resources they request. Note that RunPod's Community Cloud, Oblivus, and Cudo are all "open clouds", meaning compute is provided by third parties. +- On other providers, like Oblivus Cloud, Cudo Compute, and RunPod, users have precise control over the resources they request. Note that RunPod's Community Cloud, Oblivus, and Cudo are all "open clouds", meaning compute is provided by third parties. Lium (lium.io) is also an open cloud: independent providers on Bittensor Subnet 51 set their own per-second prices, published at https://lium.io/pricing.json. - For providers without pre-set instance configurations, we have selected configurations that are roughly equivalent to AWS's options. Generally, these configurations are good for workloads that require heavy inter-GPU communication. - Where possible, regions were set to be the west or central parts of the United States. GPU availability depends on the region. - Raw data can be found in a [csv on GitHub](https://github.com/full-stack-deep-learning/website/blob/main/docs/cloud-gpus/cloud-gpus.csv). From 69f4d9ba512d0f7ce3977ba8b05d7ca3276fe2e4 Mon Sep 17 00:00:00 2001 From: surcyf123 <114649324+surcyf123@users.noreply.github.com> Date: Tue, 22 Sep 2026 20:52:36 +0000 Subject: [PATCH 2/2] Link the Lium mention with a tagged link --- docs/cloud-gpus/index.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/cloud-gpus/index.md b/docs/cloud-gpus/index.md index 083b7749..253dd964 100644 --- a/docs/cloud-gpus/index.md +++ b/docs/cloud-gpus/index.md @@ -37,7 +37,7 @@ The table below does not include all possible configurations for all providers, - Most providers, including AWS, Azure, and Lambda, provide instances with pre-set configurations. - On GCP, any suitable machine can be connected to a configuration of GPUs. -- On other providers, like Oblivus Cloud, Cudo Compute, and RunPod, users have precise control over the resources they request. Note that RunPod's Community Cloud, Oblivus, and Cudo are all "open clouds", meaning compute is provided by third parties. Lium (lium.io) is also an open cloud: independent providers on Bittensor Subnet 51 set their own per-second prices, published at https://lium.io/pricing.json. +- On other providers, like Oblivus Cloud, Cudo Compute, and RunPod, users have precise control over the resources they request. Note that RunPod's Community Cloud, Oblivus, and Cudo are all "open clouds", meaning compute is provided by third parties. [Lium](https://lium.io/?utm_source=fullstackdeeplearning&utm_medium=directory) is also an open cloud: independent providers on Bittensor Subnet 51 set their own per-second prices, published at https://lium.io/pricing.json. - For providers without pre-set instance configurations, we have selected configurations that are roughly equivalent to AWS's options. Generally, these configurations are good for workloads that require heavy inter-GPU communication. - Where possible, regions were set to be the west or central parts of the United States. GPU availability depends on the region. - Raw data can be found in a [csv on GitHub](https://github.com/full-stack-deep-learning/website/blob/main/docs/cloud-gpus/cloud-gpus.csv).