All GPUs in Europe seem to be busy, as I wanted to spin up an Nvidia L40S to train my next splat! 🫠
So in the meantime have some juicy details of a strawberry. #3dgs
So in the meantime have some juicy details of a strawberry. #3dgs
May 18, 2026 at 1:16 PM
All GPUs in Europe seem to be busy, as I wanted to spin up an Nvidia L40S to train my next splat! 🫠
So in the meantime have some juicy details of a strawberry. #3dgs
So in the meantime have some juicy details of a strawberry. #3dgs
New GPU-based MMseqs2: 20x faster searches on a single L40S (approx. as fast as a RTX 4090) vs. a 128-core CPU. This work enables to set up a very cost-efficient ColabFold MSA GPU server. 🧵🧵
📄 www.biorxiv.org/content/10.1...
💾 mmseqs.com
🗞️ developer.nvidia.com/blog/boost-a...
📄 www.biorxiv.org/content/10.1...
💾 mmseqs.com
🗞️ developer.nvidia.com/blog/boost-a...
November 15, 2024 at 11:02 PM
New GPU-based MMseqs2: 20x faster searches on a single L40S (approx. as fast as a RTX 4090) vs. a 128-core CPU. This work enables to set up a very cost-efficient ColabFold MSA GPU server. 🧵🧵
📄 www.biorxiv.org/content/10.1...
💾 mmseqs.com
🗞️ developer.nvidia.com/blog/boost-a...
📄 www.biorxiv.org/content/10.1...
💾 mmseqs.com
🗞️ developer.nvidia.com/blog/boost-a...
alternatively, how do i convince @upcloud.com to give me an rx570 instead of nvidia l40s
January 15, 2026 at 4:26 AM
alternatively, how do i convince @upcloud.com to give me an rx570 instead of nvidia l40s
Moondream playground is back, running the latest upcoming version of the model, and now with prompt suggestions! Look at how blazingly fast it is running on an L40S...
Try it out here: moondream.ai/playground
Try it out here: moondream.ai/playground
November 27, 2024 at 4:26 AM
Moondream playground is back, running the latest upcoming version of the model, and now with prompt suggestions! Look at how blazingly fast it is running on an L40S...
Try it out here: moondream.ai/playground
Try it out here: moondream.ai/playground
we take souls in exchange for free UpCloud credits, are L40S GPU servers good enough?
August 3, 2025 at 4:04 PM
we take souls in exchange for free UpCloud credits, are L40S GPU servers good enough?
I've added LTX-Video support to the Replicate ComfyUI model:
replicate.com/fofr/any-com...
Here's an example:
replicate.com/p/6a8ydnzcqs...
15s on an L40S, out of the box.
replicate.com/fofr/any-com...
Here's an example:
replicate.com/p/6a8ydnzcqs...
15s on an L40S, out of the box.
November 26, 2024 at 5:51 PM
I've added LTX-Video support to the Replicate ComfyUI model:
replicate.com/fofr/any-com...
Here's an example:
replicate.com/p/6a8ydnzcqs...
15s on an L40S, out of the box.
replicate.com/fofr/any-com...
Here's an example:
replicate.com/p/6a8ydnzcqs...
15s on an L40S, out of the box.
Is a server with 8 L40S the right target for DwarfStar? Nope, but my friends had one, and vLLM does not support Ada NVIDIA cards with DeepSeek v4 Flash. I prompted Fable to use half-point between pipelining (for fast prefill) and TP (for faster generation).
July 10, 2026 at 7:36 AM
Is a server with 8 L40S the right target for DwarfStar? Nope, but my friends had one, and vLLM does not support Ada NVIDIA cards with DeepSeek v4 Flash. I prompted Fable to use half-point between pipelining (for fast prefill) and TP (for faster generation).
On my L40s, gpt-oss:120b gets 78.5 tok/s
October 11, 2025 at 5:19 PM
On my L40s, gpt-oss:120b gets 78.5 tok/s
InfiniteHiP: Extending Language Model Context Up to 3 Million Tokens on a Single GPU
They offload the key-value cache to host memory during inference, significantly reducing GPU memory pressure. As a result, it processes up to 3m tokens on a L40s 48GB - 3x larger.
Paper: arxiv.org/abs/2502.08910
They offload the key-value cache to host memory during inference, significantly reducing GPU memory pressure. As a result, it processes up to 3m tokens on a L40s 48GB - 3x larger.
Paper: arxiv.org/abs/2502.08910
February 15, 2025 at 4:49 PM
InfiniteHiP: Extending Language Model Context Up to 3 Million Tokens on a Single GPU
They offload the key-value cache to host memory during inference, significantly reducing GPU memory pressure. As a result, it processes up to 3m tokens on a L40s 48GB - 3x larger.
Paper: arxiv.org/abs/2502.08910
They offload the key-value cache to host memory during inference, significantly reducing GPU memory pressure. As a result, it processes up to 3m tokens on a L40s 48GB - 3x larger.
Paper: arxiv.org/abs/2502.08910
Dreame en we hebben de tijd genomen om hem te leren kennen , vriendinnen hebben er ook een . Hij dweilt ook en las hijsen vloerkleed tegenkomt tilt hij zijn dweiltjes even op .
www.mediamarkt.nl/nl/product/_...
www.mediamarkt.nl/nl/product/_...
DREAME L40s Pro Ultra | Robotstofzuiger - Met Dweilfunctie - Wit kopen? | MediaMarkt
Bestel de DREAME L40s Pro Ultra | Robotstofzuiger - Met Dweilfunctie - Wit nu bij MediaMarkt. Online of in één van onze winkels
www.mediamarkt.nl
September 29, 2026 at 11:08 AM
Dreame en we hebben de tijd genomen om hem te leren kennen , vriendinnen hebben er ook een . Hij dweilt ook en las hijsen vloerkleed tegenkomt tilt hij zijn dweiltjes even op .
www.mediamarkt.nl/nl/product/_...
www.mediamarkt.nl/nl/product/_...
What's the plural of L40S?
Nvidia L40S GPUs now up on Replicate:
replicate.com/blog/nvidia-...
We've measured that on average they're about 40% faster than our A40s (depending on the model). Switch your models from A40 to L40S to get faster outputs, cheaper.
Nvidia L40S GPUs now up on Replicate:
replicate.com/blog/nvidia-...
We've measured that on average they're about 40% faster than our A40s (depending on the model). Switch your models from A40 to L40S to get faster outputs, cheaper.
November 20, 2024 at 3:42 PM
What's the plural of L40S?
Nvidia L40S GPUs now up on Replicate:
replicate.com/blog/nvidia-...
We've measured that on average they're about 40% faster than our A40s (depending on the model). Switch your models from A40 to L40S to get faster outputs, cheaper.
Nvidia L40S GPUs now up on Replicate:
replicate.com/blog/nvidia-...
We've measured that on average they're about 40% faster than our A40s (depending on the model). Switch your models from A40 to L40S to get faster outputs, cheaper.
GigaIO Gryf - portable AI Supercomputer:
- 91.6 TFLOPS (FP32)
- 733 TFLOPS (FP16)
- 1,466 TFLOPS (FP8)
- AMD EPYC 7313
- 256GB RAM
- Up to 4x GPUs (e.g. Nvidia L40S)
- 246TB NVMe storage
- 100GbE + 25GbE networking
- 2.5kW PSU, rugged case
- <55 lbs, fits in carry-on
- 91.6 TFLOPS (FP32)
- 733 TFLOPS (FP16)
- 1,466 TFLOPS (FP8)
- AMD EPYC 7313
- 256GB RAM
- Up to 4x GPUs (e.g. Nvidia L40S)
- 246TB NVMe storage
- 100GbE + 25GbE networking
- 2.5kW PSU, rugged case
- <55 lbs, fits in carry-on
May 18, 2025 at 6:47 PM
GigaIO Gryf - portable AI Supercomputer:
- 91.6 TFLOPS (FP32)
- 733 TFLOPS (FP16)
- 1,466 TFLOPS (FP8)
- AMD EPYC 7313
- 256GB RAM
- Up to 4x GPUs (e.g. Nvidia L40S)
- 246TB NVMe storage
- 100GbE + 25GbE networking
- 2.5kW PSU, rugged case
- <55 lbs, fits in carry-on
- 91.6 TFLOPS (FP32)
- 733 TFLOPS (FP16)
- 1,466 TFLOPS (FP8)
- AMD EPYC 7313
- 256GB RAM
- Up to 4x GPUs (e.g. Nvidia L40S)
- 246TB NVMe storage
- 100GbE + 25GbE networking
- 2.5kW PSU, rugged case
- <55 lbs, fits in carry-on
aqui tem 2x Xeon 32C, 512GB RAM, NVIDIA L40S 48GB e 2x1.92TB NVMe
May 6, 2026 at 5:32 PM
aqui tem 2x Xeon 32C, 512GB RAM, NVIDIA L40S 48GB e 2x1.92TB NVMe
All the open-source flux tools are now on the `any-comfyui-workflow` Replicate model.
replicate.com/fofr/any-com...
- flux canny dev (+lora)
- flux depth dev (+lora)
- flux redux dev
Here's a working example:
replicate.com/p/9q93d0f1bh...
This model also now runs on a faster L40S GPU.
replicate.com/fofr/any-com...
- flux canny dev (+lora)
- flux depth dev (+lora)
- flux redux dev
Here's a working example:
replicate.com/p/9q93d0f1bh...
This model also now runs on a faster L40S GPU.
fofr/any-comfyui-workflow – Replicate
Run any ComfyUI workflow. Guide: https://github.com/fofr/cog-comfyui
replicate.com
November 22, 2024 at 10:28 PM
All the open-source flux tools are now on the `any-comfyui-workflow` Replicate model.
replicate.com/fofr/any-com...
- flux canny dev (+lora)
- flux depth dev (+lora)
- flux redux dev
Here's a working example:
replicate.com/p/9q93d0f1bh...
This model also now runs on a faster L40S GPU.
replicate.com/fofr/any-com...
- flux canny dev (+lora)
- flux depth dev (+lora)
- flux redux dev
Here's a working example:
replicate.com/p/9q93d0f1bh...
This model also now runs on a faster L40S GPU.
For folks who are using my consistent-character model, I've updated the hardware from A40 to L40S.
It takes 60s to make 5 images, instead of 3 minutes.
Costing ~$0.06 instead of ~$0.13
Faster and cheaper 😅
replicate.com/fofr/consist...
It takes 60s to make 5 images, instead of 3 minutes.
Costing ~$0.06 instead of ~$0.13
Faster and cheaper 😅
replicate.com/fofr/consist...
fofr/consistent-character – Replicate
Create images of a given character in different poses
replicate.com
November 28, 2024 at 12:05 PM
For folks who are using my consistent-character model, I've updated the hardware from A40 to L40S.
It takes 60s to make 5 images, instead of 3 minutes.
Costing ~$0.06 instead of ~$0.13
Faster and cheaper 😅
replicate.com/fofr/consist...
It takes 60s to make 5 images, instead of 3 minutes.
Costing ~$0.06 instead of ~$0.13
Faster and cheaper 😅
replicate.com/fofr/consist...
L40s is an nvidia datacenter card, i have two of them for ML workloads.
Im just using ollama run X —verbose for testing, nothing too rigorous
Im just using ollama run X —verbose for testing, nothing too rigorous
October 11, 2025 at 6:13 PM
L40s is an nvidia datacenter card, i have two of them for ML workloads.
Im just using ollama run X —verbose for testing, nothing too rigorous
Im just using ollama run X —verbose for testing, nothing too rigorous
NVIDIA NVIDIA L40S 900-2G133-0080-000
NVIDIA L40S 900-2G133-0080-000(NVIDIA)
nttxstore.jp
February 14, 2024 at 12:31 PM
🏠 HOME DEAL
[Prime] $449.99 | dreame L40s Ultra Robot Vacuum and Mop: $449.99!
Snag it now: #ad #deal
[Prime] $449.99 | dreame L40s Ultra Robot Vacuum and Mop: $449.99!
Snag it now: #ad #deal
[Prime] $449.99 | dreame L40s Ultra Robot Vacuum and Mop, 19,000Pa Suction & Auto Empty at Amazon
dealalert.net
September 28, 2026 at 4:51 PM
🏠 HOME DEAL
[Prime] $449.99 | dreame L40s Ultra Robot Vacuum and Mop: $449.99!
Snag it now: #ad #deal
[Prime] $449.99 | dreame L40s Ultra Robot Vacuum and Mop: $449.99!
Snag it now: #ad #deal
dreame L40s Ultra Robot Vacuum and Mop, 19,000Pa Suction & Auto Empty only $449.99 (Reg $699.99)
amazon.com/dp/B0DYSWW2LN
amazon.com/dp/B0DYSWW2LN
September 28, 2026 at 4:01 PM
dreame L40s Ultra Robot Vacuum and Mop, 19,000Pa Suction & Auto Empty only $449.99 (Reg $699.99)
amazon.com/dp/B0DYSWW2LN
amazon.com/dp/B0DYSWW2LN
🔥 Aspirateur robot laveur Dreame L40S Pro Ultra White ( 19k Pa, MopExtend, DuoBrush,Lavage eau chaude & Séchage)
479 €
#BonPlan #HighTech #Promo
479 €
#BonPlan #HighTech #Promo
Aspirateur robot laveur Dreame L40S Pro Ultra White ( 19k Pa, MopExtend, DuoBrush,Lavage eau chaude & Séchage)
479 €
www.dealabs.com
September 28, 2026 at 6:41 PM
In honour of spooky month, share a 4 word horror story that only someone in your profession would understand.
l40s: is that plural?
l40s: is that plural?
Threw away the pellet
In honour of spooky month, share a 4 word horror story that only someone in your profession would understand.
rm -rf ~/
rm -rf ~/
October 12, 2025 at 2:53 PM
In honour of spooky month, share a 4 word horror story that only someone in your profession would understand.
l40s: is that plural?
l40s: is that plural?
Its an L40s, its a tensor processing card based on the 4090 silicon I believe
March 10, 2024 at 8:10 AM
Its an L40s, its a tensor processing card based on the 4090 silicon I believe
