This collaborative space allows users to contribute additional information, tips, and insights to enhance the original deal post. Feel free to share your knowledge and help fellow shoppers make informed decisions.
Deal History includes data from multiple reputable stores, such as Best Buy, Target, and Walmart. The lowest price among stores for a given day is selected as the "Sale Price".
Sale Price does not include sale prices at Amazon unless a deal was posted by a community member.
No. It has about 1/5 the FLOPS and 1/3 the memory bandwidth of a 5090. It seems to be inference-focused, but even then, there are better ways to do that. It picks a half-assed fight with both unified memory systems and discrete GPUs and loses both. Maybe I'm missing something though.
No. It has about 1/5 the FLOPS and 1/3 the memory bandwidth of a 5090. It seems to be inference-focused, but even then, there are better ways to do that. It picks a half-assed fight with both unified memory systems and discrete GPUs and loses both. Maybe I'm missing something though.
But it also is less than 1/3 of the price of the 5090 with the same amount of RAM, so a machine with three of these can load bigger or less quantized models, improving quality of output.
But Intel has a long way to go in terms of software support for AI so you should verify your use cases before committing.
No. It has about 1/5 the FLOPS and 1/3 the memory bandwidth of a 5090. It seems to be inference-focused, but even then, there are better ways to do that. It picks a half-assed fight with both unified memory systems and discrete GPUs and loses both. Maybe I'm missing something though.
General performance for gaming is about the same as a 16GB RTX 5060ti though a fair bit worse in some games. Weaker than an RTX 5070 and for gaming totally incapable of utilizing its' 32GB configuration.
For AI Intel's missing a ton of compatibility, and performance is generally just leagues behind with few exceptions. At $1299 you can get the Radeon AI Pro R9700 which also has 32GB and pretty-much spot-on double the performance for games as well as much better AI performance especially when you factor how good CUDA emulation has come along with ROCm. Still a fair bit behind Nvidia's best, but actually quite a bit closer to a fair price compared to the Arc Pro B70.
No.
The prompt processing speed in online benchmarks for Qwen3.6-27b is like 240t/s. You'll be waiting for like a minute for it to start printing the first token when the context/chat is close to its 32GB capacity (which is around 120k depending on a quant). My 128gb Macbook has about the same prompt processing speed.
It's because this GPUs VRAM bandwidth is quite low.
Also, the software stack is not robust. You won't be able to just run anything new that comes out sometimes. You'll have to wait until the model is supported by Intel's stack.
Up until recently they didn't even allow turning the fans off in Linux (once they kicked in, they used to stay on). I believe they still don't allow setting the fan curves.
Also, I believe Intel did a rug-pull with SR-IOV support, which was the Arc Pro's redeeming quality and the reason many bought from this lineup.
Overall, it can work, but prepare to be disappointed a lot.
But it also is less than 1/3 of the price of the 5090 with the same amount of RAM, so a machine with three of these can load bigger or less quantized models, improving quality of output.
But Intel has a long way to go in terms of software support for AI so you should verify your use cases before committing.
Truth
Like
Helpful
Funny
Not helpful
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
14 Comments
Sign up for a Slickdeals account to remove this ad.
But Intel has a long way to go in terms of software support for AI so you should verify your use cases before committing.
Sign up for a Slickdeals account to remove this ad.
For AI Intel's missing a ton of compatibility, and performance is generally just leagues behind with few exceptions. At $1299 you can get the Radeon AI Pro R9700 which also has 32GB and pretty-much spot-on double the performance for games as well as much better AI performance especially when you factor how good CUDA emulation has come along with ROCm. Still a fair bit behind Nvidia's best, but actually quite a bit closer to a fair price compared to the Arc Pro B70.
The prompt processing speed in online benchmarks for Qwen3.6-27b is like 240t/s. You'll be waiting for like a minute for it to start printing the first token when the context/chat is close to its 32GB capacity (which is around 120k depending on a quant). My 128gb Macbook has about the same prompt processing speed.
It's because this GPUs VRAM bandwidth is quite low.
Also, the software stack is not robust. You won't be able to just run anything new that comes out sometimes. You'll have to wait until the model is supported by Intel's stack.
Up until recently they didn't even allow turning the fans off in Linux (once they kicked in, they used to stay on). I believe they still don't allow setting the fan curves.
Also, I believe Intel did a rug-pull with SR-IOV support, which was the Arc Pro's redeeming quality and the reason many bought from this lineup.
Overall, it can work, but prepare to be disappointed a lot.
Yeah, I think Nvidia haven't even had any drops since then.
The last helicopter out of Saigon.
But Intel has a long way to go in terms of software support for AI so you should verify your use cases before committing.
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!