AMD Ryzen 5 3600 spotted in UserBenchMark, GeekBench

[zaphod159]

Really tempted to upgrade my 1700X, but probably not worth it right?

[Xlucine]

I keep seeing this, and it's a bullrubbishrubbishrubbishrubbish figure to quote tbh. The i9-9980XE, whilst being a very expensive and powerful high core count CPU, does not score that amazingly in single threaded benchmarks. It's just a sensational sounding quote. It would make much more sense to compare single threaded scores against a 9700k or 9900k.

It's noteworthy given the clocks on the 3600 - as it's the lowest clocked of the bunch, high end ryzens will be 10% higher.

Other way of looking at it is that the 3600 is about half the price of either of those CPUs, so the real comparison will be with low-end intel (with the 3600 higher clocked than the 9400, and only slightly slower in boost speed than the 9500). The 9980XE is faster than the 9400, but only a few percent short of where it should be in hexus pifast tests going by the clockspeed difference (9.8% higher speed, 6% higher score, so 97% of the IPC) - assuming pifast is accurate for modern CPUs, of course

[Korrorra]

Vs my 1700 most everything has a 10-20 percent increase. While other things the 1700 still has a lead in. Good enough till the 3700 benchmarks can be tested against.

[zaphod159]

Hmm so it might be worth an upgrade after all! But might also rather put the money into getting a new case. Ill wait to see how 3700 is for gaming I guess

[darcotech]

How it compares against Ryzen 5 1600? Ideally one that is OC to the same freq as 3600. There we could see any improvements.

[PC-LAD]

How it compares against Ryzen 5 1600? Ideally, one that is OC to the same freq as 3600. There we could see any improvements.

I think a clock per clock performance test is easier, Stick em both at 4Ghz and there you can see IPC gains (if that particular 1600 is nicely binned anyway)

[globalwarning]

How it compares against Ryzen 5 1600? Ideally, one that is OC to the same freq as 3600. There we could see any improvements.

I think a clock per clock performance test is easier, Stick em both at 4Ghz and there you can see IPC gains (if that particular 1600 is nicely binned anyway)

Can you clarify your logic and your goals, and how it's really different than what PC-LAD said? (Neither comment was particularly clear, TBH.)

It seems like there's two or three things to look at: IPCs differences between gens (ie, single core at the same frequency, and the actual frequency wouldn't matter), potential single thread performance (ie, single core tests but either both at the native or both at OCed clocks), overall performance differences (ie, all cores, either both at stock or both at OCed, again, to be fair). One would also need to be fair about RAM speeds across the board, noting that RAM OC stability on first gen Ryzen mobos was pretty bad in many cases, and a huge pain, but RAM speed boosts provide substantive performance gains.

Anything more ambiguous in testing setups isn't particularly useful.