3 ms·
I don't think building a threaded cpu that switches threads to hide memory latency would be philosophically incompatible with VLIW. I think the later generatio
by nullc 4y ago
I don't think building a threaded cpu that switches threads to hide memory latency would be philosophically incompatible with VLIW. I think the later generation itaniums were threaded too. VLIW might make it hard to share functional units, but don't-- the relatively simplicity of the core should let you have more copies of the whole shabang. The SMT would just be for hiding memory accesses not for higher functional unit utilization.
On x86 SMT gets a bad rap because its never been tremendously effective. It's much more effective in IBM's implementation on Power. It wouldn't be hard to imagine some Itanium SMT monster.
- yosefk 4y agoHow many threads would it take to hide 100ns of DRAM access latency and what serial performance would we get?
- nullc 4y agoIt doesn't have to be hidden completely, there are still caches. With 2 unstalled threads you effectively halve it (in terms of throughput), 4 unstalled threads you effectively quarter it, etc.
- Kon-Peki 4y ago> It wouldn't be hard to imagine some Itanium SMT monster. Itanium did have SMT