Top
Best
New

Posted by neomindryan 1 day ago

Running Gemma 4 26B at 5 tokens/sec on a 13-year-old Xeon with no GPU(www.neomindlabs.com)
317 points | 204 commentspage 2
ColdStream 23 hours ago|
Vaguely related. Running an LLM on a Pentium 4. Nick named NetburstGPT. Yes, it is very slow!

https://www.youtube.com/watch?v=ILV-eu90te8

throwawayffffas 1 day ago||
The 5.2 tokens per second generation is not that bad, what kills it is the 16.2 prompt processing that makes this too slow to consider even if you have the hardware lying around.
thomasjb 1 day ago||
I was inspired by that post also, got a Qwen Coder 1.5B up to 27tok/s prompt eval and 13tok/s decode on an e5-2650v2 inside a GNOME box
rhema 1 day ago||
I love my little dual core X99 board with Xeon E5 2673 V3. It's not power efficient, but I just leave it in my basement for local Jupyter Notebook stuff. Much faster than everything cloud-based for a reasonably price at my scale.
mmastrac 1 day ago||
Is it just me or does this post not mention how much RAM they had? I would love to know - I have a dual-Xeon 1U screamer with 96GB of DDR4 RDIMM just sitting around...

FWIW I'm getting a hardware max of 20 tok/s (approx topping out the GPU's compute) on my custom local diffusiongemma port running on an M3.

neomindryan 1 day ago|
hey, I’m the author. That box has 384gb, but loading the model “only” uses about 80gb.
fouc 1 day ago||
any reason you went with q8 over q4? I'm wondering if q4 would run noticeably faster or not.
neomindryan 20 hours ago|||
I think I was just following along with the previous post about running Gemma on a Xeon. Next I’m going to see which model can give the highest tokens/sec
superkuh 1 day ago||||
Such a system is RAM bandwidth limited and not compute limited Switching to q4 from q8 would decrease the amount of data needing to be loaded by half. The token generation rate would nearly double. But generally if you can do q6 or q8 and you have enough RAM you really should. Even if it's slower.
giantrobot 1 day ago|||
Token generation is nominally bandwidth limited. Prefill/prompt processing is nominally compute limited.

For CPU inference on old hardware I don't think q4 offers any benefit over q8 since the AVX unit doesn't support such small floats. I don't even think AVX supports 4-bit int math. IIRC AVX2 does.

ChrisArchitect 1 day ago||
Related:

A 10 year old Xeon is all you need

https://news.ycombinator.com/item?id=48353348

TacticalCoder 1 day ago|
Yes and a 10 year old Xeon is going to be a v4 (not a v2 as in TFA) and it's going to have DDR4 ECC, not DDR3 ECC.

I've got a 14 cores / 28 threads Xeon from 2015 that I use as a server at home (ZFS / VMs).

It's really a sweet machine.

For ricing I've got a semi-recent AMD 7700X / DDR5 RAM (from 2023 ?) which is my main machine but the real deal is my old and trusty 10 years old Xeon server.

DDR4 ECC is pricey too atm but a 10 years old Xeon is basically free now.

A 20 cores / 40 threads costs maybe 20 USD (for just the CPU). Slap that in a $100 old HP Z440 workstation and you're good to go for quite a few workloads.

Mine is only on when I'm at my computer: it's not turned on 24/7 but more like 8/7 so the entire "but it consumes energy" point is moot.

aniwalunj 1 day ago||
Truly amazing. This gives a peek into the future for what's possible.
robotswantdata 1 day ago||
Need to run this on my Xeons with AMX
simonw 1 day ago||
How much RAM did this need?
okokwhatever 1 day ago|
To me context means everything. Tokens per second is a great metric but in the real world context window is the deal breaker when a real use case is on the table.
dofm 1 day ago|
Gemma 4 26B is capable up to 256k or 262k, can't remember which.

Whether the writer's setup affects that choice I don't know.

More comments...