Ryzen 5 3600 annihilates the 2600 in leaked benchmarks

mongeese

Posts: 643   +123
Staff
Highly anticipated: We’re still a month away from 3rd-gen Ryzen’s July 7 release date, yet leakers have already spotted two Ryzen 5 3600s tested on Userbenchmark and GeekBench. While neither of these benchmarks is particularly indicative of gaming performance, they provide intriguing insight and strong cause for optimism.

The closest comparison for the Ryzen 5 3600 is the 2600, which is also a six-core, twelve thread chip that’s just 300 MHz slower with a 3.9 GHz boost clock. Across the four benchmarks, the 3600 outperforms it by a staggering 25%, which is higher than expected.

Intel’s i5-9400F which is equivalent performance wise to the 2600 and has recently dropped to $150, is beaten by just 5% in single-core benchmarks, making it a good value. But because it's only six-core, six thread chips, they’re beaten by a shocking 44% in the two multi-core tests, throwing it out of the ring.

CPU MSRP / Street GeekBench Single Core GeekBench Multi-Core Userbench Single Core Userbench Multi-Core
Ryzen 5 3600 (1) $199 100% 100% 100% 100%
Ryzen 5 3600 (2) $199 99% 96% 86% 99%
Ryzen 5 2600 $199 / $145 81% 75% 84% 83%
Intel i5-9400 $182 / $150 99% 78% 92% 61%
Intel i7-9700K $374 / $399 112% 105% 109% 96%

Courtesy of Userbenchmark, we also know the average clock speeds during the Userbenchmark tests: 3.75 GHz and 4.05 GHz. Both are below the 4.2 GHz rated boost clock, which means a significant amount of performance is still being left on the table. Additionally, the 3600 is more or less guaranteed to overclock to 4.4 GHz as that’s the clock speed of the 3600X, and there's potential to reach closer and overclock to the 4.6 GHz of the 3900X.

Of the benchmarks, all but one was conducted on current gen X470 motherboards. The GeekBench run that was conducted on an X570 actually performed worse in the multi-core test (but better in the single core), alleviating fears that Zen 2 would underperform on older generations of motherboards. Except for first gen boards, that is, which may or may not support the new processors depending on motherboard vendors.

All this raises a bizarre question though. If the 3600 performs so well, why didn’t AMD reveal any benchmarks at Computex? They were fairly forthcoming about the 3900X and 3800X, confirming the latter could match the i9-9900K in PUBG, for example. Perhaps AMD is saving the surprise for their Next Horizon E3 announcement.

Image Credit: AMD_Robert on Reddit

Permalink to story.

 
Correction: the single core results were at 4,2ghz, not 3,9ghz. And still only 1% higher than i5 9400f wich was at 4ghz, wich means it still wont beat Intel in gaming, because Intel clocks higher and it seems the IPC is basically the same. Plus adding to that there are CCX latencies.
 
Correction: the single core results were at 4,2ghz, not 3,9ghz. And still only 1% higher than i5 9400f wich was at 4ghz, wich means it still wont beat Intel in gaming, because Intel clocks higher and it seems the IPC is basically the same. Plus adding to that there are CCX latencies.

9400 boosts to 4.1GHz. The Ryzen 5 3600 was slightly faster in the benches and it has a max boost of 4.2GHz. As the article points out, it wasn't boosting that high.

As the article observes, these were tested on the older chipsets. There is another small uptick in performance on an X570 board. This goes for these tests and other leaks.

Secondly the Ryzen tests were run with just 2666MHz DDR4, whereas this article doesn't state the exact source of the 9400 bench. Only presumably averaged from user tests on the databases. I would say that is likely to mean that score is created using on average faster than 2666MHz memory.....

Everything here points to Zen 2's IPC for single threading being at least as fast as Coffee Lake. From everything that I have pored over, from all the leaks and benches I have seen, it's at least on par with Coffee Lake. Dare I say it the latest tests on an X570 have suggested that it may actually -whisper it- have slightly better single threaded IPC than Intel's best. To the tune of 5 percent.

We'll soon know for certain. But in any case, the gap will be so small it won't matter. It'll be smaller than it has ever been, to the point it'll be irrelevant if the AMD part destroys that much on multithreading.

You say 'the intel clocks higher' but it doesn't does it? The 9400 can't overclock. That is all you get from it.

Whereas the Ryzen 5 3600 will be unlocked, and potentially I think we're looking at 4.5GHz all core if we go by what the boost speeds are on the higher end parts.

You're talking 8700k performance across the board for $200 when overclocked, at a power draw of maybe 100 watts. That would be....fantastic. This is the reality coming to roost. The vast majority of people would be extremely satisfied with 8700k level performance. For half the price.
 
Last edited:
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)
Not quite following...

A 3600 @3.9 is 1-8% faster than the 9400f at 4.1 GHz, single core.

It's also 9-12% slower than a 9700k running at 4.9 GHz, single core.

Those look pretty good from here. Definitely competitive with Coffee Lake.
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)

Hunh..?
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)
Not quite following...

A 3600 @3.9 is 1-8% faster than the 9400f at 4.1 GHz, single core.

It's also 9-12% slower than a 9700k running at 4.9 GHz, single core.

Those look pretty good from here. Definitely competitive with Coffee Lake.

I was talking about multithread numbers only and a good multithread comparisson only.
See that I replied to myself about the single thread numbers which are not impressive at all given that the 3600 isn't at 3.9 as you are saying (3.9 is the 2600 clock) but at 4.05 according to the article (maybe bad soft reading as it should be 4.2?). WIth "not impressive at all" I mean precisely that is comparable to Coffe Lake (which I also mention in my reply to myself) so, while a good improvement over old Ryzen arquitecture, not in the whole picture (especially given that coffe lake is just another Skylake refresh, a 2015 tech...).
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)

Hunh..?

You didn't understand my comment?
That's ok...
 
Rich guy, eh?

Most people make purchasing choices primarily based on price. Even Intel owners.
I'm poor and have said as much in prior posts. I'm also cynical af and like to troll so my posts can be baity, but I'm harmless and it's never mean spirited or even remotely personal to any individual. There's a good chance I may get a Zen2 chip, but I'm keeping my expectations as low as possible because that's just how I roll. In a nutshell, I'm actually pretty excited and hopeful, but feigning the contrary for funsies to keep myself in check from getting too hyped.
 
Aah too bad it's compared against CPUs with different number of threads. What will be useful is comparing it against an 8700 that has the same number of cores/threads, and even if it's not the same niche we can know how it stack against the current Intel arquitecture. And/or against an 7800x that has a lower turbo speed (but it's Skylake)...

And those single core numbers aren't very good looking anyway in comparisson with those 2 Intel processors (even at the supposed lower than turbo speed). I don't see what gives much optimism unless one is only expecting not much more than near parity between the old Intel arquitectrure and the new AMD one...
But is good, of course, they improved the thing at least that amount. Now let's see how they compare in the most important thing in the world: games (haha)
Performance parity at half the price and more threads is nothing important to you but seeing 200 FPS is the most important thing in the world, no wonder why Intel has been laughing at us for over a decade.
 
10% lower clock speed for the 2600, 15% more IPC for the 3600 - shouldn't that result be exactly what is expected?
 
Intel's new 9900ks will be the gamers choice, guaranteed. AMD will only be chosen over intel by individuals who make purchasing choices largely/primarily based on price.

Nuff said... And if we are still on the "AMD offers better price vs performance" nothing changed then, it was already like that. While Intel still delivers more raw performance. The "are you rich?" is not an argument. Some people have a job and wants to spend on hardware and have the best, instead of spending on cigarettes or alcohol, for example. Saving extra 150€-200€ to have the best doesnt seem that hard. Like I said thats not an argument.

Plus, AMD confirmed the 8 core CPUs will have 2 CCX of 4 cores, so good luck with latencies in games. And this is why AMD only shown gaming graphs compared to 2700x :)
 
Last edited:
This actually means they are slower than non HT Intel with cripled cache in single core whlist having higher clock /AMD/ and having significant perfomance hit due to the new design introducing higher I/O latencies when running multicore tasks. Ryzen 1xxx,2xxx were doing better in multicore, this one is actually worse and underperfoming.

Hype is over. Amd is still bad.
 
Last edited:
Back