3 ms·
Nvidia was forced to do that because it was so early, they had no customers except PhDs. I see no evidence AMD wants to do that right now, and instead focusing
by cherioo 3y ago
Nvidia was forced to do that because it was so early, they had no customers except PhDs. I see no evidence AMD wants to do that right now, and instead focusing on extracting value from deep pocket enterprise customers.
The way things go, I think the AMD consumer card experience will only get better once AMD manage to gimp consumer cards’ ML throughput or RAM.
- DaiPlusPlus 3y ago> The way things go, I think the AMD consumer card experience will only get better once AMD manage to gimp consumer cards’ ML throughput or RAM. Que? Making things worse will make things better?
- sjsdaiuasgdia 3y agoThey'll let you do some things on a consumer card as soon as they can make sure that you can't effectively use the consumer card in place of an enterprise card.
- throwaway48476 3y ago[flagged]
- wmf 3y agoNvidia has their own market segmentation tactics, like a 24GB 4090 is $2,000 and the 48GB version is $7,000.
- DaiPlusPlus 3y agoWhy not buy 2x24GB cards and use SLI with CUDA? Or does that represent a bottleneck-so-bad that it's a dealkiller and worth the $3000 premium?
- wmf 3y agoSLI no longer exists.
- imtringued 3y agoLlama.cpp wasn't optimized for multi GPU setups. You need to split individual layers onto the GPU instead of simple offloading.
- DaiPlusPlus 3y agoso, artificial market segmentation? ...that's exactly how you get disruptive competition.
- sjsdaiuasgdia 3y agoIt's worth noting this is a game nvidia has already been playing for a long time by intentionally limiting the performance of higher precision floating point operations on consumer cards.
- throwaway48476 3y agoIf you ask ML engineers now 100% will have started on desktop cards. Trying to squeeze the most money out of everyone just makes the ecosystem fail. Look at IBM, they sell technologies that no one uses until they get hired to work on them specifically. It's profitable enough but it's not a way to grow the ecosystem/market.
- brucethemoose2 3y agoI think you are preaching to the choir, and AMD is not listening. AMD would be selling 48GB 7900s or AI-only W7900s if they really wanted a consumer card ramp. They don't. Not because they can't (they literally prevent OEMs from doing so, who would double up VRAM in a heartbeat without AMD lifting a finger), but because AMD doesn't want that.