Question about PCI-E lanes on X299

[zornyan]

You think Intel knows or cares how to make any sense of that whole PCI-e lane mess?
Plan of X299 was just to separate people from their moneys.

Just like in keeping desktop platform so long limited to four cores while mostly fiddling with iGPU no enthusiast uses for really anything...


And what was power consumption of that thing?
Likely competing with FX-9590 everyone laughed at.

power consumption should actually be better than ryzen

https://hothardware.com/

as we've seen countless times, at equal voltages the r7s use just about the same power as the 5960x/6900k, at 4.5ghz on all cores the 7820x only needs about 1.15v (since Intel's process is far more refined than amds) so it should actually be lower than amds.

people quote the tdp, but forget that Intel gives a 'worst case scenario' under heavy load with avx instructions being used, where as amd quote a 'light/medium load' to make their figures look better.

to be fair though, you need to compare the 10 core 7900x at 5ghz (that was shown) to the 16 core amd, as the Intel will be equal to the 16 core due to each core being 35% faster than the amd (if and that's a big if the 16 core threadripper clocks to 4ghz)

 

[Panos]

power consumption should actually be better than ryzen

https://hothardware.com/

as we've seen countless times, at equal voltages the r7s use just about the same power as the 5960x/6900k, at 4.5ghz on all cores the 7820x only needs about 1.15v (since Intel's process is far more refined than amds) so it should actually be lower than amds.

people quote the tdp, but forget that Intel gives a 'worst case scenario' under heavy load with avx instructions being used, where as amd quote a 'light/medium load' to make their figures look better.

to be fair though, you need to compare the 10 core 7900x at 5ghz (that was shown) to the 16 core amd, as the Intel will be equal to the 16 core due to each core being 35% faster than the amd (if and that's a big if the 16 core threadripper clocks to 4ghz)

You do forget that the Skylake 6700K @ 4.5Ghz overclock, burns almost 30% more power under 100% load, than the Ryzen 1800X @ 4Ghz, while the latter has much faster MT perf.
The Skylake-X core isn't that far away in efficiency than the Skylake core.

 

[EsaT]

people quote the tdp, but forget that Intel gives a 'worst case scenario' under heavy load with avx instructions being used, where as amd quote a 'light/medium load' to make their figures look better.

7700K seems to exceed its TDP just fine when fully stressed:
http://www.tomshardware.com/reviews/amd-ryzen-7-1700-cpu-review,5009-8.html

Really it was Intel who first got "creative" with TDP when Pentium 4 failure couldn't compete against Athlon64.
Core 2 then brought them back to competition allowing use of honest TDP.