Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

PyTorch 1.8, with AMD ROCm support

github.com

Post reply on HN
Print view View original

31–40 of 147 posts

Re: PyTorch 1.8, with AMD ROCm support

#31

Post by my123 »

NavinF wrote: post #14
rfoo wrote: post #9

Earlier quoted context omitted.

I believe TensorFlow is a top-tier deep learning framework, and it had ROCm support since 2018. > edge TPU's are absolutely top-notch for performance per $ and Watt right now Do you mean "aren't"? The performance per $ and Watt is not awesome even when it was released, I was hoping for great toolchain support but that also didn't happen.

Are there any If it exists, I wanna buy it

Jetson Xavier NX, but that comes with a high price tag. It’s much more powerful however.

Re: PyTorch 1.8, with AMD ROCm support

#32

Post by my123 »

AlphaSite wrote: post #25
NavinF wrote: post #16

Earlier quoted context omitted.

Oh oof. Thanks for saving me time not having to look up ROCm benchmarks. I find it really surprising they they don't wanna compete on performance/$ at all by not supporting consumer cards

I think this is more of an issue that they have Compute optimised and Graphics optimised cards and Vega is their last compute optimised card. It would be very nice for them to refresh their compute cards as well.

They made new compute cards, but they aren’t available to customers. (Only businesses, under the Radeon Instinct brand)

With the price to match for those...

Re: PyTorch 1.8, with AMD ROCm support

#33

Post by ipsum2 »

upbeat_general wrote: post #30
ipsum2 wrote: post #28

Earlier quoted context omitted.

What's the point? If you have enough money to buy a brand new Apple M1 laptop, you can afford a training rig or cloud credits. Any modern discrete GPU will blow away any M1 laptop for training. Is anyone training ML models on their ultra-thin laptop?

I don’t understand this logic. If someone has $1000 for an entry level m1 machine then they also have enough money for a separate rig with a GPU that’s probably another $600-1000 for something decent? Cloud GPUs are also pretty expensive. I don’t think anyone is seriously training their ML models on their ultra thin laptop but I think the ability to do so would make it easier for lots of people to get started with wh…

If someone is shelling out for a brand new, early adopter product, then they probably have a decent amount of money.

Even when TensorFlow and PyTorch implement training support on the M1, it will be useless for practically anything except training 2-3 layer models on MNIST.

So why should valuable engineering time be spent on this?

Re: PyTorch 1.8, with AMD ROCm support

#34

Post by tpetry »

ipsum2 wrote: post #33
upbeat_general wrote: post #30

Earlier quoted context omitted.

I don’t understand this logic. If someone has $1000 for an entry level m1 machine then they also have enough money for a separate rig with a GPU that’s probably another $600-1000 for something decent? Cloud GPUs are also pretty expensive. I don’t think anyone is seriously training their ML models on their ultra thin laptop but I think the ability to do so would make it easier for lots of people to get started with wh…

If someone is shelling out for a brand new, early adopter product, then they probably have a decent amount of money. Even when TensorFlow and PyTorch implement training support on the M1, it will be useless for practically anything except training 2-3 layer models on MNIST. So why should valuable engineering time be spent on this?

Because with support for M1 you can prototype your network on your local machine with „good“ performance. There are many cloud solutions etc. but for convenience nothing beats your local machine. You can use an IDE you like etc.

Re: PyTorch 1.8, with AMD ROCm support

#35

Post by dnautics »

freeqaz wrote: post #17
dnautics wrote: post #10

Earlier quoted context omitted.

Can confirm, I made an 8x rx580 rig for an r&d project specifically for that reason.

Time to re-sell into the hot GPU market now? :P

Don't work there anymore; even so that server is probably buried under loads of failed ideas.

Re: PyTorch 1.8, with AMD ROCm support

#36

Post by rfoo »

etaioinshrdlu wrote: post #15
rfoo wrote: post #9

Earlier quoted context omitted.

I believe TensorFlow is a top-tier deep learning framework, and it had ROCm support since 2018. > edge TPU's are absolutely top-notch for performance per $ and Watt right now Do you mean "aren't"? The performance per $ and Watt is not awesome even when it was released, I was hoping for great toolchain support but that also didn't happen.

Tensorflow doesn't seem to officially support ROCm, only unofficial community projects do. This is official support from PyTorch.

Tensorflow does officially support ROCm. The project was started by AMD and later upstreamed.

https://github.com/tensorflow/tensorflow/tree/master/tensorf...

https://github.com/tensorflow/tensorflow/blob/master/tensorf...

It is true that it is not Google who are distributing binaries compiled with ROCm support through PyPI (tensorflow and tensorflow-gpu is uploaded by Google, but tensorflow-rocm is uploaded by AMD). Is this what you meant by "not officially supporting"?

Re: PyTorch 1.8, with AMD ROCm support

#37

Post by dheera »

rfoo wrote: post #9
etaioinshrdlu wrote: post #3

This is pretty neat since it is the first time in years that a top-tier deep learning framework has official support for any training accelerator with open source kernel drivers. I guess the TPU also doesn't require kernel drivers because you talk to it over network instead of PCIE. But you cannot buy a TPU, only the int8 edge TPU is for sale. (And I've heard that the edge TPU's are absolutely top-notch for performan…

I believe TensorFlow is a top-tier deep learning framework, and it had ROCm support since 2018. > edge TPU's are absolutely top-notch for performance per $ and Watt right now Do you mean "aren't"? The performance per $ and Watt is not awesome even when it was released, I was hoping for great toolchain support but that also didn't happen.

Jetson Nano: 1.4 TOPS/W, Coral TPU: 2 TOPS/W ?

Of course it doesn't really help that Google refuses to release a more powerful TPU that can compete with e.g. Xavier NX or a V100 or RTX3080 so for lots of applications there isn't much of a choice but to use NVIDIA.

Re: PyTorch 1.8, with AMD ROCm support

#38

Post by carlmr »

cranium wrote: post #13
porphyra wrote: post #4

Earlier quoted context omitted.

Lack of ROCm support in consumer RDNA 2 GPUs really makes it impossible for regular people to use ROCm. As an owner of an AMD Radeon RX 6800 I'm pretty salty about it.

It's the last thing that keeps me on Nvidia with proprietary Linux drivers. I wouldn't mind ML training on a AMD card to be slower but I need my workload to be at least GPU-accelerated.

I mean I wouldn't worry too much about it, I think if something big like PyTorch supports it AMD might rethink their strategy here. They have a lot to gain by entering the compute market.

Re: PyTorch 1.8, with AMD ROCm support

#39

Post by rfoo »

tpetry wrote: post #34
ipsum2 wrote: post #33

Earlier quoted context omitted.

If someone is shelling out for a brand new, early adopter product, then they probably have a decent amount of money. Even when TensorFlow and PyTorch implement training support on the M1, it will be useless for practically anything except training 2-3 layer models on MNIST. So why should valuable engineering time be spent on this?

Because with support for M1 you can prototype your network on your local machine with „good“ performance. There are many cloud solutions etc. but for convenience nothing beats your local machine. You can use an IDE you like etc.

Because contrary to what you believe, M1 simply is not performant enough to be used to "prototype" your network. NNs can't be simply scaled up and down. It is *NOT* like those web apps which you can run on potatoes just fine as long as nobody are hitting them heavily.

Re: PyTorch 1.8, with AMD ROCm support

#40

Post by rfoo »

dheera wrote: post #37
rfoo wrote: post #9

Earlier quoted context omitted.

I believe TensorFlow is a top-tier deep learning framework, and it had ROCm support since 2018. > edge TPU's are absolutely top-notch for performance per $ and Watt right now Do you mean "aren't"? The performance per $ and Watt is not awesome even when it was released, I was hoping for great toolchain support but that also didn't happen.

Jetson Nano: 1.4 TOPS/W, Coral TPU: 2 TOPS/W ? Of course it doesn't really help that Google refuses to release a more powerful TPU that can compete with e.g. Xavier NX or a V100 or RTX3080 so for lots of applications there isn't much of a choice but to use NVIDIA.

Sorry, should have mentioned "if you have access to Shenzhen" in my post :)

What I have in mind is something like RK3399Pro, it has a proprietary NPU at roughly 3Tops / 1.5W (on paper). But its toolchain is rather hard to use. Hisilicon had similar offerings. There are also Kendryte K210 which claims 1Tops @ 0.3W but I haven't get any chance to try it.

I was already playing with RK3399Pro When Edge TPU was announced, life is tough when you had to feed your model into a blackbox "model converter" from the vendor. That's the part I hope Edge TPU excels at. But months later I was greeted by... "to use Edge TPU, you have to upload your TFLite model to our online model optimizer", which is worse!

Post reply on HN