Deliver an actually good product by using solder/good TIM

>deliver an actually good product by using solder/good TIM
>can now attain high clock speeds with good temps
>can afford to ask more for it since its shown to easily beat AMD core for core already
>basically premium price for premium product
We could've had this, it would've been perfect. Now you're either forced to buy a CPU with 10 year old singlethreaded performance, though cheap, or to pay a lot for a shoddy product that, while offering good performance, almost fries itself.
Ryzen refresh can't come soon enough, and it'd better hit 4.5Ghz.

Number of Hz doesn't mean jack. Your other points are valid however.

>Number of Hz doesn't mean jack
>when the entire reason Intel beats AMD in singlethreaded is because they clock higher
>when the entire reason a 8700k can keep up with a 1800X in multithreaded is because it clocks higher

Ah, so that's why a raspberry pi's arm processor can compete with a higher clocked system from over a decade ago. Because clocks mean something between architectures.

>Single-threaded tasks in $CURRENTYEAR

kek.

apparently matters enough that a 6 core can keep up with an 8 core

By completely blowing it's TDP rating and basically requiring you to use watercooling, sure.

by all means keep making things up user, they say it's good having a strong imagination

>6 v 8 core
Do you not understand what "single-threaded" means?

Your average air coolers can't keep a stock 8700k in check user. You either go high end air or water.

>CPU with 10 year old singlethreaded
wut? isnt it at least better than my 3930k at 4.3? its 6 years old

What's he making up? Have you seen those temps?

You could actually brew a little coffee on that fucker.

Except for one both Intel's Coffee Lake and AMD's Ryzen exclusively consist of processors with the amd64 architecture, and secondly Gookbench always gets higher scores on ARM.

ryzen single thread is on par with haswell. so roughly 4 years old.

they have comparable IPC and identical instruction sets, that's why clocks can be (roughly) compared

compare avx 256 perf between ryzen and *lake and then your hz argument stops making sense

or you could just run a 5ghz vishera cpu against a contemporary i3 that barely pushes 3.5ghz and what do you know, the FX gets fucking destroyed

I have no reason to upgrade my laptop or desktop but when I do I probably will get something from AMD as gen 1 Ryzen already offers good enough single thread performance and when the refresh in 2018 hits it'll be even better. Even if Intel's CPUs are marginally better I will buy AMD just for the sake of not giving Intel more money.

Who said anything about that PoS bench? Also try clocking a Bulldozer chip the same as say Ivy Bridge. Hell, overclock the Bulldozer to 5GHz. Clocks aren't a meaningful measure of anything aside from how many times a chip runs its cycle per second.

what's your point exactly?
>they have comparable IPC and identical instruction sets, that's why clocks can be (roughly) compared
this is what I was doing
>compare avx 256 perf between ryzen and *lake
this is what I wasn't doing

>hey user, your argument may make sense now but don't use it in a different context or it stops making sense
yeah, thanks for your input

idk I just walked in and didn't read the whole thread

assumed you were dumb, sorry user

Your assumption was correct. He doesn't know shit. All of his points are just parroting what others have said for around 5-10 years.

IPC and clockspeed influences multicore performance retard. Core count matters more, but it's not the only thing that matters. It's bizarre how Sup Forums thinks the number of cores is the ONLY thing that contributes to multicore performance. Ryzen has good IPC but its low clockspeeds are its weakness not just in single thread, but in multi as well.

You can't multi-thread certain algorithms