Ryzen 5 3600 Gaming Leaks.

[64K]

First third-party gaming benchmarks for the AMD Ryzen 5 3600 appear online

AMD’s 3rd generation desktop processors, the AMD Ryzen 3XXX series, will officially release on July 7th and it appears that the first third-party gaming benchmarks have just surfaced online. The first gaming benchmarks are for the AMD Ryzen 5 3600 CPU. Now as you may have guessed, AMD Ryzen 5...

www.dsogaming.com

 

[kapone32]

First third-party gaming benchmarks for the AMD Ryzen 5 3600 appear online

AMD’s 3rd generation desktop processors, the AMD Ryzen 3XXX series, will officially release on July 7th and it appears that the first third-party gaming benchmarks have just surfaced online. The first gaming benchmarks are for the AMD Ryzen 5 3600 CPU. Now as you may have guessed, AMD Ryzen 5...

www.dsogaming.com

Apparently this was done with a X470 MB. I will be interested to see for myself if thee is a difference.

 

[Papahyooie]

So basically 7700k single core perf, and just under 9900k multi core perf, for 200 bucks or less? I can live with that.

 

[phill]

I am looking very much forward to the 7th July.....

 

[dgianstefani]

80ns latency vs 45 for intel... still not great for gaming.

 

[phanbuey]

that's with no OC and not tweaked memory... kind of huge - if you get the latency into the 60's it will really shine.

 

[dgianstefani]

Tweaked memory makes no difference to latency, it just eliminates one type of bottleneck. The latency stems from the issue of using multiple chiplets that have to communicate with each other.

 

[dirtyferret]

So basically 7700k single core perf, and just under 9900k multi core perf, for 200 bucks or less? I can live with that.

I see stock 8700k multicore performance not 9900k...

 

[Papahyooie]

I see stock 8700k multicore performance not 9900k...

 

[dirtyferret]

View attachment 125697

I don't believe the word "just" means what you think it means

 

[Papahyooie]

I don't believe the word "just" means what you think it means

Do you really want to argue about what a subjective and relative phrase like 'just under' means....? lol.

As stupid as this conversation is, I'll bite... we're talking about gaming performance here, not cinebench. Did you even read the link? I'll give you another excerpt:

Furthermore, this particular CPU can come close to the gaming performance of the Intel Core i9 9900K when a game uses multiple CPU threads (like in Assassin’s Creed Odyssey or Final Fantasy XV).

I'd say a 6 fps difference counts as "just under" personally... Furthermore, your cinebench results don't include results for the processor in question, so I'm not really sure what you're getting at.

EDIT: And FURTHERMORE, the 2700x is LITERALLY JUST BELOW the 9900k in YOUR cinebench chart, and the Ryzen 5 3600 slots in between them in the charts in the article. You didn't think this through, did you?

I swear, you fanboys will go to any lengths...

 

[londiste]

9900K is not the right competition here. 9600K performance is what this needs to compare favorably to. In gaming, it is not far off from 9900K either..

 

[TheoneandonlyMrK]

Tweaked memory makes no difference has latency, it just eliminates one type of bottleneck. The latency stems from the issue of using multiple chiplets that have to communicate with each other.

Your talking up the latency while it has not been quantified yet, and regardless being extreme.
I would say the 2700x is just below a 9900k by that chart.

 

[Metroid]

Tweaked memory makes no difference to latency, it just eliminates one type of bottleneck. The latency stems from the issue of using multiple chiplets that have to communicate with each other.

and yet the 3600x is only a single chiplet, we will have to see how that works with 12 and 16 cores.

 

[dirtyferret]

I swear, you fanboys will go to any lengths...

fanboy?? I've yet to say anything positive or negative about the CPU other then posting benchmark scores...

but from the the other major thread on TPU about the 3600 the SB scores puts its in the 8700k/9700k range

Behold the Ryzen 3000 CPUs

www.techpowerup.com

If you don't like the scores you can continue on your hissy fit.

 

[Papahyooie]

fanboy?? I've yet to say anything positive or negative about the CPU other then posting benchmark scores...

but from the the other major thread on TPU about the 3600 the SB scores puts its in the 8700k/9700k range

Behold the Ryzen 3000 CPUs

www.techpowerup.com

If you don't like the scores you can continue on your hissy fit.

Hissy fit huh...? I did not engage you whatsoever. That's on you, killer. "I don't think just means what you think it means" ring any bells? You wanna come in and be a dick without expecting the same treatment back, I got bad news for you.

Regardless, we're talking about gaming here. My point still stands... according to the benchmarks in THIS thread, it's just below 9900k.

 

[phanbuey]

Do you really want to argue about what a subjective and relative phrase like 'just under' means....? lol.

As stupid as this conversation is, I'll bite... we're talking about gaming performance here, not cinebench. Did you even read the link? I'll give you another excerpt:



View attachment 125699

I'd say a 6 fps difference counts as "just under" personally... Furthermore, your cinebench results don't include results for the processor in question, so I'm not really sure what you're getting at.

EDIT: And FURTHERMORE, the 2700x is LITERALLY JUST BELOW the 9900k in YOUR cinebench chart, and the Ryzen 5 3600 slots in between them in the charts in the article. You didn't think this through, did you?

I swear, you fanboys will go to any lengths...

Not really -- that's with the same video card... a 6FPS difference is 7.5% which isn't small (the 2700x, is only 2fps behind). That test is run at stock with crap ram tho, so it's most likely tunable to be faster than that (maybe even faster than the 9900k @ stock).

Tweaked memory makes no difference to latency, it just eliminates one type of bottleneck. The latency stems from the issue of using multiple chiplets that have to communicate with each other.

That's 100% false. Tweaking memory directly impacts latency because latency is measured to RAM -- so yes while there is a penalty for the chiplets it's the overall ram access latency that games care about.

http://imgur.com/a/XU6na

Skylake X and Ryzen behave similarly with fast ram -- the faster the ram the lower the latency, the higher the FPS. One has IF latency the other one uses Mesh which can't keep up with RB. Ringbus just lets you use slower ram while maintaining lower latency.

If this chip has 10-15% headroom in it from memory + OC it will be a monster.

 

[95Viper]

Keep it civil.
No insulting each other.
No name calling.
Clean Discussion.

Thank You.

 

[Papahyooie]

Not really -- that's with the same video card... a 6FPS difference is 7.5% which isn't small (the 2700x, is only 2fps behind). That test is run at stock with crap ram tho, so it's most likely tunable to be faster than that (maybe even faster than the 9900k @ stock).

Fair point. I'd PERSONALLY question the usefulness of using percentages when speaking of game FPS. "Can your eyes tell the difference?" is the criteria for me. I know I can't tell a 6 fps difference unless they're being run side by side. And put in full context of price/performance, I'd say the gap gets even smaller. I definitely wouldn't pay twice the price for 6 more frames. (as I said, that's personally. I don't discount what you're saying at all.)

So I still have the same opinion: Single core speeds comparable to a 6700k for 100 bucks less? Multicore speeds comparable to a 9900k for less than half the price? Yea, I can deal with that.

Edit: @95Viper sorry... I'll drop it.

 

[Vayra86]

Fair point. I'd PERSONALLY question the usefulness of using percentages when speaking of game FPS. "Can your eyes tell the difference?" is the criteria for me. I know I can't tell a 6 fps difference unless they're being run side by side. And put in full context of price/performance, I'd say the gap gets even smaller. I definitely wouldn't pay twice the price for 6 more frames. (as I said, that's personally. I don't discount what you're saying at all.)

So I still have the same opinion: Single core speeds comparable to a 6700k for 100 bucks less? Multicore speeds comparable to a 9900k for less than half the price? Yea, I can deal with that.

Edit: @95Viper sorry... I'll drop it.

I would agree with this wrt gaming. What matters, is if the gap is noticeable. You don't spend gaming hours looking at FPS counters.

But... its more of a use case thing and it really has been since Ryzen 2xxx; if you game @ high refresh that 7-10% translates to a noticeable gap. Similarly, if you are keen on 60 FPS locked, Ryzen can also be a tad problematic in a small selection of games - the games where even Intel CPUs dive under 70 FPS, mostly. But this is niche territory; and rarely applies to newer titles too.

My conclusion from these benches wrt gaming is simple: not much has changed from past gen. If it was good for you then, it is now, and its still not quite up there with Intel's top end with an OC. But, we also haven't seen the OC (X) models yet, which I think is the more reliable number to take in terms of maxed gaming performance. As it looks from the synthetics, this gen definitely has some spare power laying around that doesn't come out too well in gaming, who knows how unlockable that is.

And overall... this is god damn Ryzen 5, boys, non X, taking shots at Intel's nearly cooking top end offering. Glorious.

 

[John Naylor]

Do you really want to argue about what a subjective and relative phrase like 'just under' means....? lol.

As stupid as this conversation is, I'll bite... we're talking about gaming performance here, not cinebench. Did you even read the link? I'll give you another excerpt:

As it read, there were two **separate** postulations put forth ... one claim about gaming performance and another separate and independent claim about multi-core performance being "just under". It's inappropriate to go back and suggest that the claim was all about gaming performance as the original wording clearly is worded in conflict with that assertion.

I don't really see any value to such discussions until we have broad reviews covering "real applications". Historically, pre-release "leaks" have been all over the place and are always cherry-picked. When I ask a user what he needs thir PC to do, no one has ever responded "run Cinebench".

I'll wait till I see benchmark performance in Adobe Premiere before I judge how well multi-core performance stacks up ... and how it scores in TPUs 22 game test suite before I judge overall gaming performance. As for multi-core application performance, I find Adobe premiere a reliable indicator as well as being one of the more prevalent reasons to have multi-core CPUs.

About 2 weeks from now when I can fill in that Adobe Premiere table and see the gaming performance in a wide spectrum of games, I'll have an opinion.

 

[Papahyooie]

As it read, there were two **separate** postulations put forth ... one claim about gaming performance and another separate and independent claim about multi-core performance being "just under". It's inappropriate to go back and suggest that the claim was all about gaming performance as the original wording clearly is worded in conflict with that assertion.

I don't really see any value to such discussions until we have broad reviews covering "real applications". Historically, pre-release "leaks" have been all over the place and are always cherry-picked. When I ask a user what he needs thir PC to do, no one has ever responded "run Cinebench".

Okay, you people are going to end up getting me banned... I'm gonna try to be nice here...

Did you read the article? My two "claims" that you are referring to, are to single core GAMING performance (games that heavily favor fast single cores) and mutli-core GAMING performance (games that are heavily threaded and favor high core count.) My two points mirror the points in the linked article in the top post EXACTLY. I do not understand how you could think that my post had anything to do with mutlicore performance outside of the context of gaming. Your bad assumptions are not my words.

 

[gronetwork]

Excellent results!

This chip is globally 48% more powerful than the famous Intel Core i7-7700K, but it also also generates 10% more performance than the Intel Core i7-8700K.

- AMD Ryzen 5 3600 vs Intel Core i7-8700K

Only the Intel Core i7-9700K is able to beat it.