4 ms·
what do you mean? I can easily run 70b on my macbook. Fits easily.
by Svoka 2y ago
what do you mean? I can easily run 70b on my macbook. Fits easily.
- diggan 2y ago> what do you mean? I can easily run 70b on my macbook. Fits easily Does your MacBook really have a 24GB VRAM consumer (GPU) card?
- rcxdude 2y agoNewer mac books have a unified memory architecture, so the GPU can use all the available memory if it wants.
- blincoln 2y agoThere's an important caveat, though. The software you're working with needs to support Metal/MPS, or you'll end up using the CPU instead of the GPU. For example, as of a few weeks ago, PyTorch still didn't support nested tensors on Metal. Probably not a deal-breaker for a lot of folks, but important to figure out before one spends thousands on hardware.
- alwayslikethis 2y agoMacs that can run it are quite a bit more expensive than a 3090. GPUs can also do finetuning and run other models with larger batch sizes which Macs would struggle with. Also, for the models that fit both, an nvidia card can run it much faster.
- seanmcdirmid 2y agoIt’s really not that expensive to get an M4 pro with 64GB. But you really want a Max or ultra, and you aren’t going to be able to do much with images. You are limited on the GPU part. 3090s aren’t sold anymore, so you have to scrounge for them in the used market, 4090s are a couple K? The local LLM scene is still pricey and confusing.
- ponector 2y ago>>not that expensive to get an M4 pro with 64GB $3900 as the starting price does not sound 'not that expensive' for me.
- seanmcdirmid 2y agoM4 Pro with 64GB of ram starts at $3900? It doesn't take that much searching to know that is BS: https://www.apple.com/shop/buy-mac/mac-mini/apple-m4-pro-chip-with-12-core-cpu-16-core-gpu-24gb-memory-512gb https://www.apple.com/shop/buy-mac/mac-mini/apple-m4-pro-chi... An upgraded M4 Pro mac mini to 64GB is $1999. You'll probably want the extra GPUs, though, bringing you up to $2199. But you probably want the max (not offered for the mini) or ultra to do serious LLM work. An M4 Max in a 14 inch MBP with 64GB of ram will cost $3900, but that's a whole step up from the M4 pro. I'm waiting to see what they do with the Mac Studio refresh (I don't need a max in a portable, and I'm betting I can get an M4 Ultra with 64GB ram for $4000 or so).
- alwayslikethis 2y agoYou're also going to want some more SSD space, because working with these model files on a 512 GB root is going to quickly become a problem. If you get the 2TB option, that's another $600.
- seanmcdirmid 2y agoThunderbolt 5 mostly takes care of the storage problem.
- ponector 2y agoYou are right. I've been looking at MacBook with M4.
- seanmcdirmid 2y agoYou would need to get an M4 Max (which is probably the right choice for real LLM work). I don't think mobile GPUs do very well on the PC side, so if you want portability, Apple might be your best bet.
- Foobar8568 2y agoI see often MBP with 48-64gb and 1TB under than 3500 CHF. Including the M4 (thx to black Friday week). Meanwhile 4090 are close to 2000CHF. I have no doubt where the actual value is.
- moffkalast 2y agoAlright but being Swiss you can probably afford to buy both and still have most of your monthly pay left over lmao.
- alwayslikethis 2y agoA 3090 often goes for $800-900 on the used market in the US. Two of these would be $1800, and you get a much more versatile machine. However, the downside is also obvious, since your two 3090s can draw up to 800 W, and there's no chance you can carry it around with you. Overall, it's not that obvious where the actual value is, as it depends a lot on what you want to use it for.
- Svoka 2y agoWhat do you mean by "versatile"? I enjoy using my macbook for literally everything I do. And it can run 70b llama with ease. Seems extremely versatile, since I will have macbook anyway. What would I do with giant tower chugging microwave worth of energy? I would better just do AWS for training. That is way more price efficient.
- tourmalinetaco 2y agoAlso much more price efficient to buy a 3090 gaming laptop. It can do everything a Macbook can, but better, and at a third of the cost. You don’t need to put all of the money you save by buying non-Apple products into GPUs, after all.
- Svoka 2y agoHonestly, not that expensive. Not sure what the problem is. It's like one first class flight from SFO to London. You know people buy those? And what would I do with 4090? Buy a tower, insert it and SSH into it from my mac to run some model? And it still won't be enough to finetune it. Much more price efficient would be to just rent some 100s in the cloud for training.
- Svoka 2y agomissed the point where I wrote something about price. If it is expensive for you - just get AWS and train your models there or 3090 I got macbook anyway and it also happily runs llama 70b.
- do_not_redeem 2y agoYes and I can fit 70b in system RAM on my linux box. But parent was talking about discrete GPUs, and shared memory architectures like Intel's integrated GPUs or Apple's unified memory will never perform as well as dedicated GPU memory.
- zozbot234 2y ago> shared memory architectures like Intel's integrated GPUs or Apple's unified memory can't compare to dedicated GPU memory. It depends. For the initial prompt+context processing, iGPUs (and NPUs) are still a win because you're compute bound at that stage. For token generation, CPU is typically faster than iGPU (or even plausibly NPU) for local quantized models because it can make better use of all available memory bandwidth, with no wasteful padding reducing your usable throughput.
- alwayslikethis 2y agoIt is slower, but I would say fairly usable for just LLM inference (still much faster than you can read). I'd say it's more of a matter of control. You can be reasonably sure that your linux box is free from ordinary forms of telemetry, which is hard to do with a Mac.