• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***

Considering your rabid defense earlier, this is quite the amusing change of heart. Sorry you don't like the pricing. However the problem seems to be your end.

Oh and comparing current sale prices of 1xxx and 2xxx Ryzen's vs 3xxx launch price is just ridiculous. If you are looking at saving a buck and don't need the advantages of a Ryzen 3xxx, you aren't in the bloody market for one are you. It's not bloody rocket science.

I understood that something is odd with this launch when we had the release timing of the Keynote. AMD can't isolate Europe, put the Keynote at 10 o'clock Chinese time in the morning on a Monday, and 7 in the evening Californian time on a Sunday.
Also, see how they don't even want to discuss very relevant evidence that the product lineup misses entries, more 12-core parts.
 
Max frequency on that is 4.3Ghz https://browser.geekbench.com/v4/cpu/13495867

ST 5868 (+3%) the score difference matches the frequency difference, roughly.


Imagine it with a 4.6Ghz ST boost, like the 3900X

It would score 6300.


looking at the oc3d 9900k review from October last year, they were getting 6457/35687. on the geekbench charts at the mo, it's more like 6200/34000. Is this taking into account a mean of the 97th% of all results, or the security vulnerability updates maybe? or just, different versions?
 
I understood that something is odd with this launch when we had the release timing of the Keynote. AMD can't isolate Europe, put the Keynote at 10 o'clock Chinese time in the morning on a Monday, and 7 in the evening Californian time on a Sunday.
Also, see how they don't even want to discuss very relevant evidence that the product lineup misses entries, more 12-core parts.
It was a bank holiday weekend in the UK, so I'm not too sure I understand your point...
 
New 3600 (4.15GHz) Geekbench with 3733 ram
http://browser.geekbench.com/v4/cpu/13631495

Single-Core Score
5716
Multi-Core Score
29990

that's at 4.15, i have seen 4.4 on other recent geekbenches

Using this score to extrapolate, if SC scales linearly with frequencies, here are my predictions for the other SKU Geekbench 4 SC scores--

3600x/3700x -- 6043
3800x-- 6180
3900x-- 6318
3950x-- 6455


https://old.reddit.com/r/Amd/comments/c44s1l/definitive_3733mhz_r5_3600_windows_10_geekbench_4/
 
Does SC scale in a linear fashion on AMD cpus? I don't think it does for Intel

being discussed on that reddit link:

reddit people said:
DwaSokoly 12 points 5 hours ago


Good extrapolation. One tip from me but I could be wrong.

If you look at the comparison of latest Ryzen clock incrementals vs. GeekBench 4.0 scores incrementals according to these clocks then tou find that the GB4.0 scores incrementals are as an average about 75% of clock incremental -> because in general CPU performance and GB4.0 performance won't scale linear. To adjust this effect you should moltiply clock incremental by 0.75 to get (I think that) more accurate scores.

And as I calculate scores for 3733MHz memory should be like that:

3600x/3700x -> 5918

3800x -> 6019

3900x -> 6119

3950x -> 6219

However as we could see from the other scores with not optimal memory settings such scores wouldn't be an average. Average scores could be as follow:

3600x/3700x -> 5500-5600

3800x -> 5600-5700

3900x -> 5700-5800

3950x -> 5800-5900

But, to be honest, these scores are also very, very good scores especially according to first Ryzen generation (+35+40% of incremental)!
 
I just found that Geekbench includes a score of your memory in the combined SC and MC scores.

That makes it really hard to compare CPU's using geekbench as you need to find tests where both cpus have the exact same RAM
 
I just found that Geekbench includes a score of your memory in the combined SC and MC scores.

That makes it really hard to compare CPU's using geekbench as you need to find tests where both cpus have the exact same RAM
It sure does, and that's why GB4 seems to be offering ridiculous levels of uplift compared to last gen. Clearly, this won't be replicated in all tasks. That being said, 50%+ extra memory bandwidth, at roughly the same latency, can't be a bad thing.
 
Im pretty much a noob to all these technical things yous are all talking about and im not understanding it lol

Is a single core better than multi core for gaming ?

Im looking to build a new mini itx pc mainly for gaming and im undecided whether to wait for these ryzens or just go intel. I have a I5 4690k and im looking for a decent upgrade to last me 5+ years

Per-core performance is still the primary metric to be looking for, but if you're wanting to keep a system for half a decade, you will want at least the 6-core option, if not the 8-core. Personally, I would persuade you into the 8-core territory.

The next-generation consoles will be using 8-core Zen 2 processors at around 3.00 GHz, which means that anybody owning such CPUs in their PC should see a tangible benefit to their overall playability since the frequency deficit will be used by the console manufacturers to ensure that all of the cores are properly taken advantage of.

This then causes a domino effect on the desktop/laptop side of things, and that 8-core purchase will be worthwhile.


ahh, ok. Thought I read when the 9900k reviews dropped that MCE only pushed an all core boost of 4.7 on the 9900k. Which I thought made sense seeing as an Auto MCE function would probably push more than 1.30Vcore which would generate more heat than most were prepared for. I'm more confused as to what a 9900KS brings to the table now then, unless it's 8086K quality binning and anothe +£100!

MCE pushes the all-core turbo to match the single-core turbo, which would be 5.00 GHz in this case. :)

With Intel dropping prices by 15%, I suspect the 9900KS might not cost as much as originally thought, but it's clear that Ryzen will be cheaper. The extra 300 MHz might help Intel keep up, however.


I'd be very surprised if they actually release the 9900KS, and if they do it will be a paper product only. They need to concentrate on getting the 10c/20t 1151 CPU out, with at least 4.6GHz boost speed all core at <250w, and maybe a 5.0GHz boost on 2 cores.

Speaking of which, does anyone remember the i9-9900KFC? Whatever happened to those rumors...?

Any chance of it resurfacing if the i9-9900KS can't get the job done? Rely on eDRAM, on top of the 5.00 GHz all-core frequency? :D


So you have no intention on actually buying a Zen2 CPU it would seem. So please don't bother posting in here if you literally have nothing to add ,other than rubbish.

That's some dedication!


They have been EOL'ing them for a while, i doubt they have been making any since they started making the IO Dies for Zen 2 at GloFlow.

They also have certain promises to fulfill with the PRO line-up for businesses. They have to be available for at least two years after introduction, and the Zen+ PRO line-up was released only in September.


These were with 2133Mhz Memory - so 3733Mhz gives one hell of a boost!

2700X
Single-Core Score
4410
Multi-Core Score
22546

3800X
Single-Core Score
5406
Multi-Core Score
34059

https://browser.geekbench.com/v4/cpu/compare/13607469?baseline=13567135

Looking forward to see the 3900X with 3733Mhz

Interesting to note that the only area where it loses is in memory latency.

Surprised nobody has picked up on that.


looking at the oc3d 9900k review from October last year, they were getting 6457/35687. on the geekbench charts at the mo, it's more like 6200/34000. Is this taking into account a mean of the 97th% of all results, or the security vulnerability updates maybe? or just, different versions?

Possibly between unapplied security mitigations, and variations in the benchmark version. Annoyingly, the article doesn't give any indication to the software versions used whatsoever.

What is clear, is that their 2700X result is 12% faster than @N19h7m4r3's, which is more than margin of error difference. They used DDR4-3200 for their testing, which should have closed that 2700X gap considerably if the benchmark version was the same.


I just found that Geekbench includes a score of your memory in the combined SC and MC scores.

That makes it really hard to compare CPU's using geekbench as you need to find tests where both cpus have the exact same RAM

And this is why I strongly dislike going by the two figures at the top.
 
Last edited:
looking at the oc3d 9900k review from October last year, they were getting 6457/35687. on the geekbench charts at the mo, it's more like 6200/34000. Is this taking into account a mean of the 97th% of all results, or the security vulnerability updates maybe? or just, different versions?

No idea, i don't normally take any notice of Geekbench.

I'm simply using it to compare performance between Zen + and Zen 2 because we are getting results of Zen 2 posted to it, out side of that i think its pretty useless.
 
looking at the oc3d 9900k review from October last year, they were getting 6457/35687. on the geekbench charts at the mo, it's more like 6200/34000. Is this taking into account a mean of the 97th% of all results, or the security vulnerability updates maybe? or just, different versions?
https://browser.geekbench.com/v4/cpu/13637164 - This sytem is fully patched.

SC - 6482 MC - 38684
 
I understood that something is odd with this launch when we had the release timing of the Keynote. AMD can't isolate Europe, put the Keynote at 10 o'clock Chinese time in the morning on a Monday, and 7 in the evening Californian time on a Sunday.
Also, see how they don't even want to discuss very relevant evidence that the product lineup misses entries, more 12-core parts.
I hope the straws you are grasping are paper and not plastic.
 
Back
Top Bottom