Intel Gaudi2 chips outperform Nvidia H100 on diffusion transformers

https://stability.ai/news/putting-the-ai-supercomputer-to-work

by memossy • 2 years ago

146 60 2 years ago

memossy11• 2 years ago

"For Stable Diffusion 3, we measured the training throughput for the 2B Multimodal Diffusion Transformer (MMDiT) architecture model. Gaudi 2 trained images 1.5x faster than the H100-80GB, and 3x faster than A100-80GB GPU’s when scaled up to 32 nodes. "

Mistletoe• 2 years ago

I can feel the NVDA stock slipping as we speak…

It has been amazing watching the groupthink at work on that stock when we just saw the same group do it on TSLA to disastrous effect. A similar no moat situation where they simply can’t imagine competitors ever existing.

varispeed• 2 years ago

Typically stocks fall once I buy them and go up after I sell them. I am not planning on buying NVDA for now, so likely it will keep going up.

* just to be clear - this is a joke

bluedino• 2 years ago

My father in law was telling me the other day that he was going buy some NVIDIA stock because it is going to go up to 1,400

memossy11• 2 years ago

It's a great company & will do well, plenty of demand & B100s/BH200s etc coming

The Hopper stuff is particulalry interesting

gitfan86• 2 years ago

The Model Y is the best selling car in the world in 2023. Those of us who were buying in 2019 are still up quite a bit even though the stock was higher at one point. RIVN, Ford, GM all are losing a lot of money on every EV they sell. We were right to bet on TSLA being a major winner.

I actually put 40% of my TSLA into NVDA last year, because the demand for AI hardware is going to keep going up. I'm not saying the stock will never go down, I'm sure it will be volatile, but don't confuse short term volatility with long term technologic transformations.

MasterScrat2• 2 years ago

Interesting! this was already the case with TPUs easily beating A100s. We sell Stable Diffusion finetuning on TPUs (dreamlook.ai), people are amazed how fast and cheap we can offer it - but there's no big secret, we just use hardware that's strictly faster and cheaper per unit of work.

I expect a new wave of "your task, but on superior hardware" services to crop up with these chips!

memossy11• 2 years ago

The v5es and v5ps are pretty amazing at running SD, giving code for SD3 now to optimise it on those.

v5es are particularly interesting given the millions that will land and the large pod sizes, particularly well constructed for million token context windows.

renewiltord• 2 years ago

Which TPUs do you use? Cloud-hosted or your own hardware? Interesting insight.

1024core2• 2 years ago

"TPUs" are a Google-only product, available* only on GCP.

* Notwithstanding the Choral boards

memossy11• 2 years ago

We use v4s, v5es & v5ps. Mostly v5ps, very stable int8 training (versus the horror that is fp8 stability)

MasterScrat2• 2 years ago

We started on V3s, now fully moved to V4s with some V5Es, investigating a full move towards V5E & V5P

doctorpangloss• 2 years ago

But you and I can't buy a TPU. You and I can buy an H100.

isoprophlex• 2 years ago

Speak for yourself! I can't even afford 1/10th of an H100.

tromp• 2 years ago

I found this Intel website [1] more informative regarding architecture and capabilities of Gaudi2:

[1] https://www.intel.com/content/www/us/en/developer/articles/t...

jsheard• 2 years ago

Hasn't H100 been shipping in volume for about a year already? Is Gaudi2 even available at comparable scale yet? I wouldn't count Nvidia out until they start slipping on similar timescales, i.e. if B100 doesn't have a clear lead over competing parts that become available at roughly the same time.

memossy11• 2 years ago

I think as we go to enterprise workloads the total cost of ownership becomes important.

NVIDIA is still the best for research given ecosystem but once the models are standardised as with transformers/LLaMA and likely multimodal diffusion transformers it then becomes about scale, availability and cost per flop.

mittermayr• 2 years ago

Frankly, this may be good to level out the market a bit. While it's been fun to see Nvidia rise up through this insanity, it would only be healthy to have others catch up here and there eventually.

ekelsen• 2 years ago

Some analysis of how and/or why it is able to be 3x faster despite no hardware metric being 3x better would make this actually useful and insightful instead of advertising.

Flux159• 2 years ago

This is nice to foster some competition in hardware for model training, but the availability of these machines seems very limited - I don't think there's any major cloud provider allowing per hour rental of Gaudi2 VMs and Intel's own site directs you to buy an 8x GPU provisioned server from Supermicro for more than 40k USD. Availability and software stack is still heavily in Nvidia's favor right now, but maybe by the end of the year that will start changing.

memossy11• 2 years ago

Think it'll probably crack on with Gaudi3 at 4x performance, twice VRAM etc later this year.

We found cuda sycl conversion surprisingly good https://www.intel.com/content/www/us/en/developer/articles/t...

machdiamonds• 2 years ago

It's hard to guess these cards' real performance uplifts. According to Nvidia, H100 is 11x faster than A100, but that's definitely not true in most cases. If Gaudi3 is legitimately 4x faster than Gaudi2, it should be a very good value proposition compared to even the B100. I'm really curious whether Intel will be able to compete with X100 using Falcon Shores or not. Regardless, I don't think Nvidia's margins are sustainable.

GaggiX• 2 years ago

Intel's own site directs you to buy an 8x GPU provisioned server from Supermicro for more than 40k USD

Isn't that the price of a single H100?

sabujp• 2 years ago

what you're telling me is that SMCI doesn't care if NVDA or INTC or even AMD outperforms on their hardware, they'll make axes and shovels for anyone and profit either way :)

GC_tris• 2 years ago

Disclaimer: I technically still am employed at Genesis Cloud (though no longer actively involved).

Genesis Cloud started integration and testing of Gaudi2 quite a while ago. I fully agree with the take of the article.

I can't promise per hour rental, but for longer times they are available! (should you be interested you can find contact details on the website)

az2262• 2 years ago

Would you rent out a node for few days for benchmark testing?

az2262• 2 years ago

Link for the price?

ABS2• 2 years ago

H100 was released almost exactly 1 year ago so I guess it's ok if Intel is now ready to compete with last year's model.

To those commenting about "no moat" remember CUDA is a huge part of it, it's actually HW+SW and both took a decade to mature, together

memossy11• 2 years ago

It took less than a day to port our code over, we do custom CUDA across modalities.

Gaudi2 was actually announced 2 years ago and is 7nm like the A100 80Gb it was meant to be competitive with, Gaudi3 later this year is probably going to be the inflection point as that ramps

The cost is like 1/3

https://www.intel.com/content/www/us/en/newsroom/news/vision...

ABS2• 2 years ago

"Announced" 2 years ago is different from its availability and ability:

- Intel acquired Habana in 2019

- Habana launched Gaudi2 in 2022

- only in H2 2023 Habana enabled FP8 which delivered around 100% improvement in time-to-train

On the rest I believe you but markets don't move based on single individual's/company's data points

nabla9• 2 years ago

  2024:  Nvidia's B100 TSMC 3nm (?)
  2024:  Intel Gaudi3  TSMC 5nm (*)
  2023:  AMD MI300X    TSMC 5nm/6nm 
  2022:  Nvidia H100   TSMC 4N
  2020   Nvidia A100   TSMC 7nm

(*): performance critical chiplets at least.

imtringued• 2 years ago

The fact that AMD's GPGPU platform is buggy for consumers has more to do with incompetence and product cannibalisation than the difficulty of building properly working drivers. Machine learning uses profoundly simple operations. Building a pytorch backend isn't difficult if the drivers are working properly.

thunderbird120• 2 years ago

Gaudi3 is supposedly due this year with a 4X bump in Bf16 training over Gaudi2. Gaudi is an interesting product. Intel seems to have something pretty decent but it hasn't seen much of a volume release yet. Maybe that comes with V3? Not sure exactly what their strategy with it is.

We do know that in 2025 it's supposed to be part of Intel's Falcon Shores HPC XPU. This essentially takes a whole bunch of HPC compute and sticks it all on the same silicon to maximize throughput and minimize latency. Thanks to their tile-based chip strategy they can have many different versions of the chip with different HPC focuses by swapping out different tiles. AI certainly seems to be a major one, but it will be interesting to see what products they come up with.

memossy11• 2 years ago

It was interesting Aurora used GPU Max & definitely looking forward to Falcon Shores.

I think Gaudi2 was bad timed & they had to build stack, Gaudi3 is where I think we will see mass adoption given availability, way cheaper price/performance & maturer stack.

There is still weird stuff when using them but they are surprisingly solid.

1024core2• 2 years ago

NVIDIA's profit margin is almost 92% on an H100. I'm surprised more chip companies haven't jumped on a "ML accelerator" bandwagon by now.

wmf2• 2 years ago

There's a dozen AI chips already; how many do you want?

Now working ones is a different story.

adtac• 2 years ago

For anyone deeply familiar with building one: what are the biggest problems you run into 3 months in that you didn't foresee?

Just curious because IME that's the point where the fun problems surface :)

BryanLegend• 2 years ago

This message was brought to you by Intel

memossy11• 2 years ago

I mean they work well, here is another blog by Databricks: https://www.databricks.com/blog/llm-training-and-inference-i...

yukIttEft• 2 years ago

I'm wondering how AI scientists work these days. Do they really hack Cudakernels or do they plug models together with highlevel toolkits like pytorch?

Considering its the latter, considering pytorch takes care of providing optimized backends for various hardwares, how big of a moat is Cuda then really?

david-gpu• 2 years ago

Pytorch relies heavily on the extensive libraries of high-performance kernels provided by NVidia, such as cuDNN.

In other words, it goes something like this:

    Application
    Pytorch (and similar)
    cuDNN (and similar)
    CUDA (and similar)
    NVidia GPU

My opinion, based on what I saw those wizards do, is that reproducing the feature set and efficiency of cuDNN/cuBLAS is deeply nontrivial.

cherryteastain2• 2 years ago

One question I have that nobody, including an Intel AXG employee, has been able to answer satisfsctorily for me is why both Gaudi and Ponte Vecchio exist. Wouldn't Intel have better chances of success if they focused on one product line?

georgeburdell• 2 years ago

Gaudi was brought into Intel via an acquisition. Ponte Vecchio was an internal program. It can be explained by a combination of management silos and perhaps pre-existing obligations for Ponte Vecchio with the government for how they both came into being

cherryteastain2• 2 years ago

Sad if it's a management/politics happenstance than a genius master plan.

wmf2• 2 years ago

It's good risk reduction, especially since Ponte Vecchio failed.

meragrin_• 2 years ago

From my understanding, Gaudi specializes in a specific use case (deep learning/AI) while Ponte Vecchio is more generic HPC. Also, DL/AI accelerators may not work well with newer models so the generic HPC hardware may be the only option for certain models until the DL/AI accelerators have a chance to catch up.

mistrial9• 2 years ago

https://es.wikipedia.org/wiki/Antoni_Gaud%C3%AD

Gaudi is a famous name for a reason.. the flowing lines and frankly, nonsense and silliness, in the art and architecture of Gaudi stands for generations as a contrast to the relentless severity of formal classical arts (and especially a contrast to Intel electronic parts).

throwaway4good2• 2 years ago

Who fabs the Gaudi2? TSMC or Intel themselves?

throwaway4good2• 2 years ago

TSMC 7nm it seems.

qeternity• 2 years ago

Has anyone been running LLMs on TPUs in prod? Curious to hear experiences.

emadm• 2 years ago

Yeah they train well and very stably even int8, maxtext now has LLaMA and mistral support too, pytorch xla gets 50% MFU with spmd and you have some nice stacks like levanter

Haven't been too impressed with inference versus tensor rt llm for example though

CrocODil• 2 years ago

Does the performance picture change with Int8?

lostmsu• 2 years ago

I would potentially be interested in Gaudi-based workstation. Supermicro servers seem good, but they do not have DisplayPort outputs, and jury-rigging them on is not something I'd do.

Intel Gaudi2 chips outperform Nvidia H100 on diffusion transformers

Related Stories