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

Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

blogs.nvidia.com

Post reply on HN
Print view View original

331–340 of 347 posts

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#331

Post by pjmlp »

fulafel wrote: post #312
pjmlp wrote: post #277

Earlier quoted context omitted.

OpenCL is only supported via hacks to install shared libraries into one own's device. Vulkan even if optional until Android 10, it is supported by the SDK since version 7, which is something that OpenCL never had. No serious Android developer would make their life even harder than it already is with the official APIs, by making use of an API that requires device owners to manually install libraries via ADB.

Hmm, is the OpenCL situation really that fringe on android? Eg this OpenCL info app[1] description says "Even though OpenCL™ isn't part of the Android platform, it's available on many recent devices. On Android it's usually used as a back-end for other frameworks like Renderscript. Some manufacturers are providing SDKs for developers to use OpenCL™ on Android. " I addition to the mentioned PowerVR and Intel platforms…

Yes it is, it is not exposed via the SDK, regardless how Renderscript ends up being compiled to machine code.

You won't find any reference to OpenCL here,

https://source.android.com/

https://developer.android.com/

So while those SDKs do exist, they aren't for application developers, rather for OEMs themselves.

You as application developer have zero control about what GPUs the customers might be using, there is no way to control it on Android manifests, only to specify what APIs are expected, which again, don't have OpenCL as part of the list.

https://developer.android.com/guide/topics/manifest/manifest...

So if you as application developer want to use said SDKs, it is only for your own device, most likely rooted, don't expect to sell applications on the store using OpenCL.

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#332

Post by fulafel »

pjmlp wrote: post #331
fulafel wrote: post #312

Earlier quoted context omitted.

Hmm, is the OpenCL situation really that fringe on android? Eg this OpenCL info app[1] description says "Even though OpenCL™ isn't part of the Android platform, it's available on many recent devices. On Android it's usually used as a back-end for other frameworks like Renderscript. Some manufacturers are providing SDKs for developers to use OpenCL™ on Android. " I addition to the mentioned PowerVR and Intel platforms…

Yes it is, it is not exposed via the SDK, regardless how Renderscript ends up being compiled to machine code. You won't find any reference to OpenCL here, https://source.android.com/ https://developer.android.com/ So while those SDKs do exist, they aren't for application developers, rather for OEMs themselves. You as application developer have zero control about what GPUs the customers might be using, there is no way…

Yep, you can't of course rely on OpenCL support, you have to provide a fallback. But are there any issues with including the SDK supplied OpenCL ICDs with the APK, for use with validated GPUs?

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#333

Post by m463 »

bredren wrote: post #203
DonHopkins wrote: post #173

Earlier quoted context omitted.

And he sure has a LOT of really nice spatulas! His Spatula City Frequent Buyer Card must have a lot of stamps on it. https://www.youtube.com/watch?v=2XbCWmY0eqY

Those kids bouncing up and down in the back seat with no seatbelts on looks like madness.

It's like the episode of Mad Men where the kids are playing spaceman: https://youtu.be/2XbCWmY0eqY

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#334

Post by sk0g »

shubuZ wrote: post #325
sk0g wrote: post #29

Earlier quoted context omitted.

Mind you, this is an optimised version of BERT, from what I could see in their blog post.

The network topology is the same, no modules are removed, etc. Of course it will be optimized to run on Ampere

I'll link the the mention of the optimised versions, but that's not what I mean!

Say earlier model XYZ trained 4 epochs per hour, and BERT trained 2 epochs per hour. Now on a single card if you can train optimised BERT for 4 epochs per hour, that doesn't necessarily mean the same card will handle XYZ at 8 epochs per hour.

It's a technological achievement nonetheless, but the fact that it was heavily optimised for the new architecture, possibly beyond an extent infeasible for non-Nvidia developers, still has to be considered.

[0] https://nvidianews.nvidia.com/news/nvidia-achieves-breakthro...

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#335

Post by rajnathani »

tbenst wrote: post #265
XCSme wrote: post #259

Earlier quoted context omitted.

I am not that much into ML, just fiddled with it a bit, is tf32=fp16?

Not quite, but close. “tf32” is 18 bits, but with the same 10 bits of exponent that fp32 has. It’s the range of fp32 with the precision of fp16. It’s a shame to see such unoriginality in new number representations. I’d much rather see Posit hardware acceleration: https://web.stanford.edu/class/ee380/Abstracts/170201-slides...

TF32 is 19 bits, not 18 bits. There's an additional bit for sign.

https://blogs.nvidia.com/blog/2020/05/14/tensorfloat-32-prec...

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#336

Post by chemmail »

blattimwind wrote: post #39
dijit wrote: post #18

Earlier quoted context omitted.

I don't like that people downvoted you for asking a question. If someone thinks the question is stupid or not doesn't mean that a downvote is warranted. (nor an upvote, answer the question and move on.) To answer though; it's just a coincidence, as you might already know Nvidia uses famous scientists (especially in the field of electricity) as the names of their microarchitectures. * Volta (Alessandro Volta, inventor…

Tesla is famous in the US for inventing polyphase AC and induction motors, but this is really one of these stories were a bunch of people invented the same thing very closely to each other due to a precipitating reaching of understanding. Note that Tesla's designs were IIRC two-phase which is largely inferior to three-phase. The push for three-phase and associated designs and inventions (three phase transformers on a…

If rebuilding a grid today, would going to DC power be advantageous to AC?

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#337

Post by beyondcows »

chemmail wrote: post #336
blattimwind wrote: post #39

Earlier quoted context omitted.

Tesla is famous in the US for inventing polyphase AC and induction motors, but this is really one of these stories were a bunch of people invented the same thing very closely to each other due to a precipitating reaching of understanding. Note that Tesla's designs were IIRC two-phase which is largely inferior to three-phase. The push for three-phase and associated designs and inventions (three phase transformers on a…

If rebuilding a grid today, would going to DC power be advantageous to AC?

What has changed between then and now? High voltage DC also works well for long distance transmission but is only preferable to AC at even higher voltages.

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#338

Post by blattimwind »

beyondcows wrote: post #337
chemmail wrote: post #336

Earlier quoted context omitted.

If rebuilding a grid today, would going to DC power be advantageous to AC?

What has changed between then and now? High voltage DC also works well for long distance transmission but is only preferable to AC at even higher voltages.

HVDC works for linking grids and for very long distance transmission between few stations. It does not work well at all for distribution. Never did. Never will. Sorrynotsorry.

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#339

Post by pjmlp »

fulafel wrote: post #332
pjmlp wrote: post #331

Earlier quoted context omitted.

Yes it is, it is not exposed via the SDK, regardless how Renderscript ends up being compiled to machine code. You won't find any reference to OpenCL here, https://source.android.com/ https://developer.android.com/ So while those SDKs do exist, they aren't for application developers, rather for OEMs themselves. You as application developer have zero control about what GPUs the customers might be using, there is no way…

Yep, you can't of course rely on OpenCL support, you have to provide a fallback. But are there any issues with including the SDK supplied OpenCL ICDs with the APK, for use with validated GPUs?

Yes, that is not how Android works.

You are not allowed to ship drivers like that, and since version 7, there is kernel validation about stable NDK APIs.

So yeah just because there is an OpenCL SDK for Mali, doesn't mean that a random Android device with Mali will have the drivers or kernel support in place, because that isn't something that Android requires.

Google has collaborated with Adobe in porting their OpenCL shaders to Vulkan.

If they actually cared, they would have made OpenCL available instead.

Re: Nvidia CEO Introduces Nvidia Ampere Architecture, Nvidia A100 GPU

#340

Post by fulafel »

pjmlp wrote: post #339
fulafel wrote: post #332

Earlier quoted context omitted.

Yep, you can't of course rely on OpenCL support, you have to provide a fallback. But are there any issues with including the SDK supplied OpenCL ICDs with the APK, for use with validated GPUs?

Yes, that is not how Android works. You are not allowed to ship drivers like that, and since version 7, there is kernel validation about stable NDK APIs. So yeah just because there is an OpenCL SDK for Mali, doesn't mean that a random Android device with Mali will have the drivers or kernel support in place, because that isn't something that Android requires. Google has collaborated with Adobe in porting their OpenCL…

I see, thanks for the explanation. Oh well, let's see where things go with Vulkan compute.
Post reply on HN