- Aug 26, 2014
You're not factoring in that a 6/12 SKU will undoubtedly be clocked higher than an 8/8 SKU, which should easily make up for any performance deficit in the "virtual" threads. This is especially important as most workloads today outside of video rendering or other very parallel workloads have varying performance requirements per thread (and mostly require 1-2 high performance threads and a bunch of low-to-mid priority background tasks., and ultimately, single thread performance is still the most important. If a 6/12 SKU is clocked 2-300 MHz higher than an 8/8 SKU (which, given the same TDP, seems likely), it would be better for the vast majority of consumer workloads.Let's see how this "worse performance" looks like depending on CPU load, considering SMT would account for about +25% IPC increase and taking 6c/6t as the base for relative performance increase:
While some may argue SMT offers a bit more power efficiency than extra raw cores - leading to higher clocks , this will only happen in scenarios where thread count well past 8, more likely 10-12.
- Until 6 threads both 8c/8t and 6c/12t will perform at almost identical performance levels with a 6c/6t, including power usage.
- At 7 threads the 8c/8t will offer 16% more performance and the 6c/12t will bring +4% over 6c/6t.
- At 8 threads the 8c/8t will be at +33% , the 6c/12t will offer +8%.
- At 10 threads the 6c/12t will jump at 16% relative advantage, still considerably behind the 33% advantage of the "inferior" 8c/8t.
- It takes 12 threads for the 6c/12t to hit it's max throughtput of +25%, and with the help of somewhat smaller power usage at full load come out even or slightly on top of the 8c/8t.
Moral of the story? SMT needs higher thread count to truly shine. Physical cores kick in faster.
No they haven't. Where are the HEDT i5s, if I might ask? They don't exist, as they have long since realized that at high core counts, non-SMT SKUs make little to no sense.Ask Intel, they've been doing this kind of wasting for many years now.
Here you're partially right. I see 6/12 as a good compromise - 4/8 will undoubtedly win in therms of ST performance, but as games and applications grow ever more parallel, two more physical cores (and four more threads) will become increasingly more important. And compared to 8/8 or 8/16, the clock deficit compared to 4/8 wouldn't be nearly as large for 6/12. Intel's HEDT line show this pretty clearly - when moving from 6 to 8 cores you only lose 100MHz max boost, but a whopping 400MHz in base clocks (which do matter more in high usage scenarios). This is clearly to conserve power/lower heat output.If that were true, then consumer would have nothing to gain from a 6c/12t chip either. Everything above 4c/8t simply doesn't make any sense, high clocked Intel 4/8 chips also beat lower clocked 6/12 chips.
How would this maximize their revenue? You realise that an 8/8 SKU would require them to take fully functional chips, disable SMT, and sell them at a lower price, right?Well, yeah. The point of it is to maximize AMD's revenue.