AMD Ryzen 5 3600 benchmarked early, shows performance barely slower than Intel i9-9900K

[Cal Jeffrey]

Why it matters: An early review of AMD’s Ryzen 5 3600 has leaked and shows the processor outperforming Ryzen 7 2700X in several benchmarks and lagging only slightly behind Intel's ninth-generation Core i9 in productivity workloads. Official reviews (including TechSpot's) are coming next month.

Spain-based tech reviews site El Chapuzas Informatico put the newest AMD CPU to the test against the Ryzen 7 and an Intel Core i9-9900K and found it running between the two and even beating both in some cases. This seems to confirm a previous leak we reported on at the beginning of the month.

The Ryzen 5 3600 was set up on a Gigabyte Aorus Gaming 7 WiFi motherboard with G.Skill FlareX DDR4@3200 MHz, and a GeForce RTX 2080 Ti Founders Edition GPU. In Cinebench R15 and R20 runs, the Ryzen 5 3600 beat the Zen+ based 2700X in single-core performance. It did not fair as well for multi-core, but that is no surprise considering it is 6-core/12-thread versus 8C/16T.

Even with a higher memory latency of 80.5ns, the Ryzen 5's performance was not hampered to a significant degree. For example, in wPrime 2.10 32M the 3600 almost kept pace with the Core i9-9900K falling only slightly behind. It was faster than previous generation CPUs including Intel Core i7-8700K, 8600K, and AMD Ryzen 7 2700X and 1700X. In the multi-core wPrime tests, the Ryzen 7 2700X slightly outperformed its next-gen counterpart.

It was noted that the Ryzen 5’s memory write speed was unusually low. It only hit 25.6GB/s, while the 2700X and 1700X were over 47GB/s. “However this could be due to the motherboard being used or maybe an issue with sub timings on the memory, something that will need to be verified in future reviews,” TechPowerUp speculates.

All in all, AMD’s Ryzen 5 3600 could not keep up with the ninth-generation Intel Core processor, but it didn’t lag far behind. As gains against previous AMD offerings, the 3600 showed it is a step up, for the most part, especially considering its lower clock speed and fewer cores/threads.

For an MSRP of $200, the Ryzen 5 3600 should make a decent choice for mid-range gaming rigs, especially if it overclocks well. El Chapuzas Informático has some gaming benchmarks for Assassin's Creed Odyssey, Far Cry 5, Final Fantasy XV, and Total War: Warhammer 2 if you're interested in those. The 3600 ran right behind the i9 in all of those.

As with all unofficial news, hold off your excitement until something more concrete comes out. AMD’s Ryzen 5 3600 drops on July 7, and you can bet we'll have our full review ready in due time.

All benchmarks graphs by El Chapuzas Informatico

Permalink to story.

https://www.techspot.com/news/80685-amd-ryzen-5-3600-benchmarks-leak-showing-only.html

 

[QuantumPhysics]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

 

[Vrmithrax]

Similar (or slightly slower) performance for half the price of the i9... Interesting for sure, it'll be fun to see if AMD gains more traction with power users.

 

[Vulcanproject]

Far Cry 5 was the worst gaming result. Possibly an optimisation issue.

The other 3 gaming tests the 3600 was ~5 percent behind a 9900k. Not bad, considering it is running only 4.2GHz maximum boost and the 9900k manages 5GHz. This is the bottom end, I'll emphasise, the very bottom end of this Zen 2 range on launch.

Even if you can only bang it to 4.5GHz all core then you have a very fast chip for the money. No problem matching an 8700k/9600k by the looks of it, and probably duffing them both up good in multi core and power consumption wise.

Now show me what the 3700x can do with 8 cores and 4.4GHz boost. Plus an overclock....

 

[TheBigT42]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Ryzen 5 3600 MSRP $200
Intel i9-9900K Street Price of ~$500

Which one would you buy?

 

[jbc029]

Would have beem nice to see it run up against a 9600k, since that is the price commensurate part.

 

[crazypostalbob]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Ryzen 5 3600 MSRP $200
Intel i9-9900K Street Price of ~$500

Which one would you buy?

To be fair, the equivalent would be much more like a i5 9400 vs the Ryzen 5 3600, which the 9400 is $199 right now.

As to which I would buy in that category, the Ryzen, of course. The performance would likely be very similar at least in what I care about (gaming) I would be very pleased if that single core improvement actually makes it overall in gaming better (we have far too *edit* little benchmarks to know right now).

Even if it were slightly slower I would still get it on principle. Intel got out of control with pricing and had next to zero major changes for years. 4 core chip, 4 core chip 2, 4 core chip 3, 4 core chip redux 4, I'm not exaggerating. They were really stagnant. Ryzen made the trend more cores, and a downward overall price per core.

Intel 2500k ($216), 3570k (230), 4670k ($230) , 6600k ($243)

All these were 4 core chips.

Now the intel 9400 6 core is $199. That's quite a shake up, but it was far greater a shake up in the higher core markets. Those also trended up in the 8 core, and this is the biggest difference.

6 core (no 8 core out at this time) 4960x $999 release price, 8 core 5960x $1,050, 8 core 6900k $1,089 (there was no X series 8 core, so I used the K), 8 core 7820x (8 core) $589. You read that right, they cut the price of the 8 core in half. I wonder what happened around the 7820x release date to cause this cut in June of 2017's intel release?

It could have been the February release of the Ryzen 1700 and 1800x.

This alone makes me want to buy one, even though to date my hardcore gamer won't let me. My fight the man won't let me get an intel cpu. It's a struggle eh?

 

[Nobina]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Ryzen 5 3600 MSRP $200
Intel i9-9900K Street Price of ~$500

Which one would you buy?

He'd buy i9 for $1000 if it was 1% faster because it's still faster. That's how modern people think.

 

[Adi6293]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

You dont have to be the best if you price your product right

 

[wiyosaya]

Wake me on July 7.

 

[Lionvibez]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

You make it sounds like its so simple.
This is still David vs Goliath, intel still has alot more money and can afford to launch and fail with products AMD cannot.
Not to mention how many years and billions does it take to design a cpu?

 

[EEatGDL]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Almost for sure next year with Zen 3. Intel has nothing new for the next years in desktop and, take this with a grain of salt, Zen 3 would supposedly introduce 4 threads per core. Also, TSMC's 7nm+ would be a 30% higher density than 7nm, so see it as a 6nm.

 

[pcnthuziast]

The prices people are expecting are unrealistic. Based on this performance and the current market, I expect price parity with something like an 8700k, ie well over 300 dollars.

 

[Evernessince]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Ryzen 5 3600 MSRP $200
Intel i9-9900K Street Price of ~$500

Which one would you buy?

It's not worth the time on Quantum, who frequently says things along these lines:

""If you think I'm bad now, you should have seen me in my Best Buy days back during college where I would absolutely disparage against specific products and push what I felt was the best."

- QuantumPhysics, May 26th, 2019"

The prices people are expecting are unrealistic. Based on this performance and the current market, I expect price parity with something like an 8700k, ie well over 300 dollars.

The pricing used is directly from AMD. Don't get more realistic then that.

 

[pit1209]

The prices people are expecting are unrealistic. Based on this performance and the current market, I expect price parity with something like an 8700k, ie well over 300 dollars.

The thing is that this are official prices, the 3600 is going to cost $199, about the performance then we have to wait and see.

 

[Aaron Fox]

Are we going to play the game of "let's pretend there aren't any serious Intel-only security vulnerabilities and therefore ignore the performance regressions required to fully deal with them?"

Apple isn't the only company to say that hyperthreading has to be DISABLED for full security on Intel processors. I just read an article yesterday lamenting the way Intel rolled out a lot of CPUs, with not very many having hyperthreading. Hmm... I wonder why. The writer said "Intel is using it as a premium feature". Or, maybe Intel was doing some damage control.

The tech press seems to be falling down in its responsibility to hold companies like Intel adequately accountable for their bad design decisions. Everyone is expected to pretend that security doesn't matter and that if it does everyone has free money to replace their chips with defective design. Does Intel even have a CPU out right now that fixes its hyperthreading breach? I don't think so.

And yet, here we are, again and again, with articles talking about performance that don't even mention "security", "hyperthreading", or anything of the sort — as if none of any of it exists. Get real please.

And, before anyone chimes in to try to downplay the problem... it, at the very least, deserves to be mentioned in EVERY performance head-to-head with AMD, so the uniformed can, at least, know that there are potentially serious performance deficits should they choose to care about security. Hyperthreading is just one of the most egregious.

 

[Aaron Fox]

Are we going to play the game of "let's pretend there aren't any serious Intel-only security vulnerabilities and therefore ignore the performance regressions required to fully deal with them?"

Apple isn't the only company to say that hyperthreading has to be DISABLED for full security on Intel processors. I just read an article yesterday lamenting the way Intel rolled out a lot of CPUs, with not very many having hyperthreading. Hmm... I wonder why. The writer said "Intel is using it as a premium feature". Or, maybe Intel was doing some damage control.

The tech press seems to be falling down in its responsibility to hold companies like Intel adequately accountable for their bad design decisions. Everyone is expected to pretend that security doesn't matter and that if it does everyone has free money to replace their chips with defective design. Does Intel even have a CPU out right now that fixes its hyperthreading breach? I don't think so.

And yet, here we are, again and again, with articles talking about performance that don't even mention "security", "hyperthreading", or anything of the sort — as if none of any of it exists. Get real please.

And, before anyone chimes in to try to downplay the problem... it, at the very least, deserves to be mentioned in EVERY performance head-to-head with AMD, so the uniformed can, at least, know that there are potentially serious performance deficits should they choose to care about security. Hyperthreading is just one of the most egregious.

If any of these performance regressions caused by kludging around Intel's design defects affect single-threaded performance it should be mentioned and fully covered in this article.

In fact, even if none of them do it should be explicitly mentioned so people will know the answer.

It's inexcusable to ignore this serious problem. There is more happening in computing than Cinebench and cruddy video games.

 

[Evernessince]

Are we going to play the game of "let's pretend there aren't any serious Intel-only security vulnerabilities and therefore ignore the performance regressions required to fully deal with them?"

Apple isn't the only company to say that hyperthreading has to be DISABLED for full security on Intel processors. I just read an article yesterday lamenting the way Intel rolled out a lot of CPUs, with not very many having hyperthreading. Hmm... I wonder why. The writer said "Intel is using it as a premium feature". Or, maybe Intel was doing some damage control.

The tech press seems to be falling down in its responsibility to hold companies like Intel adequately accountable for their bad design decisions. Everyone is expected to pretend that security doesn't matter and that if it does everyone has free money to replace their chips with defective design. Does Intel even have a CPU out right now that fixes its hyperthreading breach? I don't think so.

And yet, here we are, again and again, with articles talking about performance that don't even mention "security", "hyperthreading", or anything of the sort — as if none of any of it exists. Get real please.

And, before anyone chimes in to try to downplay the problem... it, at the very least, deserves to be mentioned in EVERY performance head-to-head with AMD, so the uniformed can, at least, know that there are potentially serious performance deficits should they choose to care about security. Hyperthreading is just one of the most egregious.

Unfortunately most people don't care about security until it's already too late.

 

[Dragonstongue]

So, you are comparing effectively an i5 AMD style, to an i9 and saying "the i9 is still faster, though not in everything"

hmm, so, 2700x "top end" for Ryzen+
3600 is "nearly keeping up" and is basically at the 1/2 point (core count, clock speed, using far slower memory than are capable of etc etc etc.

IMO if they are "as good" as leak suggest, and near matching Intel's current "best they can do" which is core i9/i7 PERIOD.

point is...unlikes other generations, is cool to see them take one notch lower product which is more or less "parity" (not to mention....WATTAGE.....3600 likely cost less than 2700x and WAY LESS than i9 yada yada....TDP 65w vs 105 vs 125-180 (depend on which 9xxx chip)

IMO...that is like giving you that race inspired V6 that keeps up with most V8 though only using power of and cost of ~V4 with turbos....

How can anyone NOT see this as "DUDE that is a good deal IF they release at least as good as appears"

AMD is likely sandbagging results to keep Nv/Intel in dark as much as possible till the 7th (all it take is them to have a master driver they give to OEM etc, day of launch)

AMD not stupid (or dead long ago) and Dr Lisa Su is A HELL of a CEO let alone doctorate at very very high level....smart as a whip I bet...she took AMD out of a double digit billions in the hole to not 3 years later very profitable, higher ASP per product, much less cost to produce them than Nv or Intel is doing current.

Ryzen 3xxx and Navi are going to make many "shut up and listen" (if they got a brain between their ears and not just green/blue colored slime in there ^.^

 

[Aaron Fox]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Considering the fact that the tech press likes to ignore — that AMD chips don't have to have hyperthreading DISABLED for full security... there is already a massive performance advantage.

That's just one of the many Intel-only vulnerabilities with performance regressions that the tech press is happy to completely ignore, aside from tepid articles that are only about that if you're fortunate enough to even get those. Some sites don't even bother to do the minimum, as with Extremetech that never wrote an article about, for example, the Spoiler vulnerability.

Meltdown. RIDL. ZombieLoad. Fallout. MDS. Spoiler. etc. etc. etc. etc.

"AMD has reported only 15 security vulnerabilities until now. However, we can see a total of 233 vulnerabilities on Intel’s listing."

Given how many Intel-only security problems that come with regressions have been exposed in recent months it seems safe to assume there are more waiting to come to the fore in short order. But, whenever there is a performance head-to-head these days, people act like it's back before any of them were made public and don't even mention the word security.

 

[Aaron Fox]

Unfortunately most people don't care about security until it's already too late.

The tech press isn't 'most people'. It's comprised of professionals whose job it is to be better than the engagement and depth level of the average computer user.

So, that excuse won't pass muster. It's not acceptable to do any performance head-to-head without bringing up the security performance regression issue.

 

[ShadowDeath]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Lol... You are aware that the 3600 is their mid ranged CPU right? They have faster.....

 

[Xxtankmasterx]

So when exactly is AMD going to release something that blows everyone away and isn't woefully behind Intel and Nvidia because from what I can tell, it ain't easy catching up to the lead.

Come on now man. This is a low-midrange Ryzen vs the flagship consumer processor from Intel. the chip is probably running on an engineering board, and the chip itself may be an engineering sample. The Ryzen 5 3600 will also be priced between half and a third that of the i9

 

[Bp968]

The prices people are expecting are unrealistic. Based on this performance and the current market, I expect price parity with something like an 8700k, ie well over 300 dollars.

What do you mean "expecting"? AMD announced pricing at computex and E3. The chip being discussed here is 199$. It is known.

 

[Manuel Diego]

the chip is probably running on an engineering board, and the chip itself may be an engineering sample.

Or probably not. I mean, read the article, it's an X470 board and a final product they used. No need to exagerate...