chapstickbomber

Reducing supply adds further value?

Lmao

That's what I said in 2016 😮‍💨

Zen4 is just faster Zen3 mostly, so fast Zen3 leaves little room for gain, your shit is zooming.

What is pupil swim?

I had a 5800X, went to AM5 on a 7950X3D, then I last week put a 5950X in the old board for some compute and can manual OC via Ryzen Master to 4.3GHz and the 7950 only hits 4.6GHz at stock with like 10% more performance, so it's mostly clocks and cache stuff.

5950X is a fuckin beast.

To be fair "highest selling high end GPU ever" for AMD is still not a lot compared to NV or their own midrange stuff.

Navi21 wasn't high end, GA102 was just weak, also 6900 XT was not at 1.19% in late 2021,that's 🧢

Which is hilarious because 7900 XT(X) is the highest selling high end GPU they've ever made afaik

ASRock AQUA 7900 XTX is actually the fastest, highest power model this gen. Each partner puts more or less effort into each model they put out. Sometimes you get an extruded aluminum brick and a piddly VRM, sometimes you get 2kg bricks and 1000A of power.

very good presentation

Another brute force solution that can improve stability and gaming consistency out of the box is to lower the max clock of CCD1 cores in BIOS so that CCD0 cores always clock higher. Without any special service running, Windows will often pin high load work to the "faster" cores which is one reason why 7950X3D benchmarks are all over the place. This means a zero or 1 percent hit in heavy MT, and maybe a 5-6% hit in procs that scale perfectly with frequency and were natively being assigned to CCD1 cores hitting 5.6 or 5.7GHz instead of 5.3 on CCD0. Re: that 6%: YAGNI tbh

5800X runs hot as fuck because it can't even hit the full 142W stock TDP on non-goat cooling.