Intel's next-gen Arrow Lake CPUs might come without hyperthreaded cores

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by juggertao on Tuesday January 23, 2024 @01:34AM (#39100253)

Yet GPUs which take HT to the next level by having thousands of "hyper-threads" work very well for scientific computing.

[ Parent | on HN ]

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by pjmlp on Tuesday January 23, 2024 @04:29AM (#39101285)

Their design is completly different, it isn't shared execution units like it happens on CPUs.

[ Parent | on HN ]

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by juggertao on Tuesday January 23, 2024 @05:03AM (#39101491)

But it is. GPUs have many more threads in flight than execution units.

[ Parent | on HN ]

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by pjmlp on Tuesday January 23, 2024 @05:46AM (#39101758)

Threads groups get exclusive resources in SIMT execution pipelines.

[ Parent | on HN ]

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by juggertao on Tuesday January 23, 2024 @06:11AM (#39101902)

And at memory stall they are exchanged with other waiting thread groups.

Just like HT.

[ Parent | on HN ]

Re:Intel's next-gen Arrow Lake CPUs might come without hyper... by RealHusky on Wednesday January 24, 2024 @07:14AM (#39116379)

Scheduling algorithm is different.

CPUs target low latency (they switch often). GPUs target high troughput (they switch rarely, only when needed).

High troughput algorithms dont have problem with a lot of threads. Low latency algorithms have problem with a lot of threads (they need lot of cache memory because of constant switching).

[ Parent | on HN ]