You'd still need a fairly large amount of compute power to be able to run DeepSeek R1 locally, no?
Of course this is for a personal instance, you'd need a much more expensive setup to handle concurrent users. And that's to run it, not train it.
"hello how are you today?" - 7 tokens.
And this is so much better than I could have imagined in a very short span of time.
This takes advantage of the sparsity of MOE and the efficient KV-cache of MLA.
E-ATX case = ~$300
Power Supply= ~$300
Xeon W-3500 (8 channel memory) = $1339 - $5889
Memory = $300-$500 per 64GB DDR5 RDIMM
Memory will be the major cost. The rest will be around $5,000. A lot less than "$100,000"!
And the original message you were responding to was using a CPU with AMX and mixing it with a GPU like Nvidia 4900/5900. That way the large part of the model sits in the larger slower memory, and the active part in the GPU with the faster memory. Very cost effective and fast. (Something like generating 16 Tokens/s of 671B Deepseek R1 with a total hardware cost of $10-$20k.) They tried both single and dual CPU, with the latter about 30% faster....not necessarily worth it.
https://github.com/kvcache-ai/ktransformers/blob/main/doc/en...
That's the theory. In practice, Sapphire Rapids needs 24-28 cores to hit the 200 GB/s mark and it doesn't go much further than that. Intel CPU design generally has a hard time saturating the memory bandwidth so it remains to be seen if they managed to fix this but I wouldn't hold my breath. 200 GB/s is not much. My dual-socket Skylake system hits ~140 GB/s and it's quite slow for larger LLMs.
> Why does it have to be a dual CPU design?
Because memory bandwidth is one of the most important limiting (compute) factors for larger models inference. With dual-socket design you're essentially doubling the available bandwidth.
> And the original message you were responding to was using a CPU with AMX and mixing it with a GPU like Nvidia 4900/5900.
Dual-socket CPU that costs $10k on a server that costs probably couple of factors more. Now you claimed that it doesn't have to be that expensive but I beg to differ - you still need $20k-$30k of worth equipment to run it. That's a lot and not quite "cost effective".
(Dual Socket Skylake? Do you mean Cascade Lake?)
If you price it out, it's basically the most cost effective set-up with reasonable speed for large (more than 300 GB) models. Dual socket basically doubles the motherboard[2] and CPU cost, so maybe another $3k-$6k for a 30% uplift.
[1] https://www.intel.com/content/www/us/en/products/sku/231733/... $3,157
[2] https://www.serversupply.com/MOTHERBOARD/SYSTEM%20BOARD/LGA-... $1,800
Please price it out for us because I still don't see what's cost effective in a system that costs well over $10k and runs at 8 tok/s vs the dual zen4 system for $6k running at the same tok/s.
I am not sure what your point is? There are some nice dual socket Epyc examples floating around as well, that claim 6-8 tokens/s. (I think some of those are actually distilled versions with very small context sizes...I don't see any as thoroughly documented/benchmarked as the above). This is a dual socket Sapphire Rapids example with similar sized CPUs and a consumer graphics card that gives about 16 tokens/second. Sapphire Rapids CPU and MB are a bit more expensive, and a 4090 was $1500 until recently. So for a few thousand more you can double the speed. Also the prompt processing speed is waaaaay faster as well. (Something like 10x faster than the Epyc versions.)
In any case, these are all vastly cheaper approaches than trying to get enough H100s to fit the full R1 model in VRAM! A single H100 80 GB is more than $20k, and you would need many of them + server just to run R1.
The math is clear: single-socket ktransformers performance is 8.73 tok/s and it costs ~$12k to build such a rig. The same performance one gets from a $6k dual-EPYC system. It is a full-blown version of R1 and not a distilled one as you say.
Your claim about 16 tok/s is also misleading. It's a figure for 6 experts while we are comparing R1 with 8 experts against llama with 8 experts. 8 experts on dual-socket system per ktransformer benchmarks runs at 12.2 - 13.4 tok/s and not 16 tok/s.
So, ktransformers can roughly achieve 50% more in dual-socket configuration and 50% more than dual-EPYC system. This is not double as you say. And finally, the cost of such dual-socket system is ~$20k and therefore isn't the "best cost effective" solution since it is 3.5x more expensive for 50% better output.
And tbh llama.cpp is not quite optimized for pure CPU inference workloads. It has this strange "compute graph" framework which I don't understand what is it there for. It appears completely unnecessary to me. I also profiled couple of small-, mid- and large-sized models and the interesting thing was that majority of them turned out to be bottlenecked by the CPU compute on a system with 44 physical cores and 192G of RAM. I think it could do a much better job there.
Cheapest 32 core latest EPYC (9335) x 2 = $3,079.00 x 2
Intel 32 Core CPU used above x 2 = $3,157 x 2 (I would choose the Intel Xeon Gold 6530 which is going for around $2k now, and with with higher clock speeds, and a 100 MB of more cache)
AMD Epyc Dual Socket Motherboard Supermicro H13DSH = $1899
Intel Supermicro X13DEG-QT = $1,800
Memory, PSU, Case = Same
4090 GPU = $1599 - $3,000 (temporary?)
Besides the GPU cost, the rest is about the same price. You only get a deep discount with AMD setups if you use EPYCs a few years old with cheaper (and slower) DDR4.
And again, if you go single CPU, you save over $4,000, but lose around 30% in token generation.
The "$6,000" AMD examples I've seen are pretty vague on exactly what parts were used and exactly what R1 settings including context length they were run at, making true apple to apple comparisons difficult. Plus the Sapphire Rapids + GPU example is about 10x faster in prompt processing. (53 seconds to 6 seconds is no joke!)
Yes, you're blatantly misrepresenting information and moving goalposts. Right now it has become clear that you're doing this because you're obviously affiliated with ktransformers project.
$6k for 8 tok/s or $20k for 12 tok/s. People are not stupid. I rest my case here.
So, say 500W. That's, for me in my expensive electricity city, $40/million tokens, with the pretty severe rate limit of 5600 tokens/hours.
If you're in Texas, that would be closer to $10/million tokens! Now you're at the same price as GPT-4o.
Related, you can get a whole lot of cloud computing for $2k, for those same experiments, on much faster hardware.
But yes, the data stays local. And, it's fun.
This comment chain is pretty funny.