Viewing profile — neilmovva

I think Hopper's native matmul tile is 64x64, and Blackwell is 128x128. see this blog for a reference on Blackwell: https://hazyresearch.stanford.edu/blog/2025-03-15-tk-blackwe...

The multilingual example in the launch graphic has Qwen3 producing the text: > "Bonjour, pourriez-vous me dire comment se rendreà la place Tian'anmen?" translation: "Hello, could y…

Today, training in "low precision" probably means computing FP8 x FP8 -> FP32. The FP32 accumulation is still important, but otherwise yes this works, especially if we're talking a…

Not really: 5090: 210 TF / $2k == 105 TF/$k B200: 2250 TF / $40k == 56 TF/$k Getting only 2x the FLOPs per dollar probably isn't worth the hassle of having to rack 10x as many GPUs…

I was surprised to see 5090's theoretical BF16 TFLOPs at just 209.5. That's not even 10% of the server Blackwell (B200 is 2250, and GB200 is 2500). B200 costs around $30-40k per GP…

A bit surprised that they're using HBM2e, which is what Nvidia A100 (80GB) used back in 2020. But Intel is using 8 stacks here, so Gaudi 3 achieves comparable total bandwidth (3.7T…

I agree that synchronization causes overhead, so 2x GPUs won't achieve the ideal 0.5x total runtime. But here, taking your Alpaca benchmark as an example, we are seeing 2x GPUs get…

promising results, excited to try it out! question on the perf benchmarks: why do all the results with 2 GPUs & DDP take longer than the single GPU case? Both benchmarks do the sam…

I can't universally agree with the headline statement. The article focuses on the pros of SRAM, which are real -- peak bandwidth (e.g. 5 TB/s out of the H100’s L2) and lower energy…

I like this review: https://www.lighterra.com/papers/modernmicroprocessors/ A bit dated, but the major ideas used in current CPUs are all covered!

> moves to Austin because it is less “vulnerable to climate change” > commutes by plane hmmm

A bit underwhelming - H100 was announced at GTC 2022, and represented a huge stride over A100. But a year later, H100 is still not generally available at any public cloud I can fin…

Yes, we all place a lot of trust in cloud vendors today. FHE is a way to move the trust boundary back to the client - let the server be as malicious or insecure as it wants. Raw co…

Thanks for the feedback, I understand your hesitation. We don't just want to advertise guarantees - we want you to never trust third-party servers again. Fully homomorphic encrypti…

Thanks for checking it out! Responses inline: > That sounds like loading the entire database every time Yup, we do perform computation over the entire database for every read - the…

Thanks! Yup, it's not always practical to make a huge number of queries when you expect many of them to come back empty. Instead, we first perform private lookups against a Bloom f…

Thanks! Yup, private retrieval is interesting as a product because it's a fundamentally new capability; there aren't really competitors we can show incremental improvements against…

Our FHE scheme uses lots of Number Theoretic Transforms (NTTs), which are pretty computationally expensive. NTT is a good candidate for acceleration, and there is quite a bit of in…

While OPT-175B is great to have publicly available, it needs a lot more training to achieve good results. Meta trained OPT on 180B tokens, compared to 300B that GPT-3 saw. And the …

Hi HN! I'm a cofounder of Blyss.dev, a YC W23 company building an SDK for homomorphic encryption. In recent HN discussions of homomorphic encryption, there seems to be a common opi…

I was curious about exactly how much burst heat could be absorbed, so I asked WolframAlpha [0]. In a 15" workstation laptop, I think the CPU could quite reasonably pull +100 watts …

Latest in a trend of silicon industry consolidation. A few other major moves in the embedded market over the last five years: NXP + Freescale in 2015 Microchip + Atmel in 2016 ON S…

The author's comments on cache sizes are a bit reductive. Not all "L3" is created equal, and designers always make tradeoffs between capacity and latency. In particular, the EPYC p…

Actually, the L3 cache is also sharded across chiplets, so there's a small (~8MB) local portion of L3 that is fast, while remote slices will have to go over AMD's interdie connecti…