Preliminary Tests Of Intel Sandy Bridge & Ivy Bridge vs. Broadwell

[curaga]

Oh, indeed there's a very tiny difference. NVM.

[newwen]

The per-watt performance is quite good.

But considering that Skylake is coming in the second half of this year, I would recommend anybody to hold back till the back-to-school season and get a skylake-based system with DDR4.

[Guest]

Meh, I was considering upgrade but these numbers haven't impressed me all that much (efficiency is nice, but since performance matters to me the most, I don't see any real benefits per $ spent).
If power consumption remained the same and performance doubled over 3 generations of CPUs, then I would consider giving intell all my money and waifu

[l8gravely]

Happy to run more tests with whatever new test profiles or PTS patches are provided by the community.

Maybe I'll try and see about adding more tests to get perf/watt measurements on linux based tests. Nothing quick though... too many other demands on my time.

[mendieta]

What's the big deal? Really?

I really wonder why is everybody going 'no big deal':

1. For mobile processors, considerably less power at modestly better performance is huge: you either get better battery life or a lighter device with similar battery life.
2. For desktops, where performance is limited by thermal issues (plus ipc), better ipc and much better performance per watt should mean room for higher raw performance at similar TDP (probably, a lot more room for overclocking)
3. In both cases, graphics are 10 to 25% faster than Haswell, depending on the test. This is not yet 'discreet card performance', but it makes more games playable at decent resolutions, and already, meany are with Haswell.

So, enough b!tching, folks