AMD Ryzen 7 5800X3D Geekbenched, About 9% Faster Than 5800X

[Bwaze]

I have a feeling if this is that good in geekbench, it will be a monster in games.

But that's the thing - it isn't.

9% increase just in multicore and even slightly lower score in single core load usually means better productivity (rendering, vidro encoding and other stuff that scales well), not better gaming.

 

[Dr. Dro]

Gauge with anything but geekbench!

Wanna bet its an engineering sample? We dont know any facts, nor what build its in, so shouldnt be taken as fact, wait for the real reviews.

I find it extremely unlikely, I would argue that it is not an engineering sample processor. It was finalized quite a while ago. All AMD engineering samples follow a strict nomenclature, for example:

AMD Eng Sample: <OPN> followed by processor type, characteristics, whether it is unlocked or not, etc.

To the best of my knowledge, unlike Intel they do not ship out to any partners late ES or qualification sample processors with a final CPUID designation programmed into them and even then, programs like CPU-Z and Geekbench would be able to detect they are engineering samples in the event that any would ship with final designation, because they would have the ES bit set.

I don't find it very surprising that many applications simply do not show marked improvements over the regular 5800X, they share a common architecture and topology, after all. Geekbench was one of those I expected the highest improvements, though, since it has quite a few subtests that would highly benefit from the cache - it seems there is one subtest that it seems to be at least 15% ahead of the 5800X though.

 

[Punkenjoy]

But that's the thing - it isn't.

9% increase just in multicore and even slightly lower score in single core load usually means better productivity (rendering, vidro encoding and other stuff that scales well), not better gaming.

Not all single core results are equals.

 

[Ruru]

But that's the thing - it isn't.

9% increase just in multicore and even slightly lower score in single core load usually means better productivity (rendering, vidro encoding and other stuff that scales well), not better gaming.

Though modern games do not rely on single thread anymore.

 

[Valantar]

But that's the thing - it isn't.

9% increase just in multicore and even slightly lower score in single core load usually means better productivity (rendering, vidro encoding and other stuff that scales well), not better gaming.

But again: the workloads tested are not representative of gaming CPU workloads, regardless of the number of cores used. That's the whole problem with using GB (or other canned CPU benchmarks) for gauging gaming performance - correlation does not imply causation, and while there might be similarities across the performance characteristics in gaming and some workloads, those do not necessarily apply across architectures, clock speeds, cache amounts, etc. So while you're right in principle that "ST performance has more in common with gaming performance than MT performance", that's about as indicative of gaming performance as four legs generally being faster than two for running - which fails to mention rather crucial aspects such as the physiology of the animal, its size, etc. A cheetah and a sloth both have four limbs used for locomotion, yet the speed difference is quite dramatically different.

 

[Maelwyse]

Faster?

Here's a comparison with a bog standard non-OC'ed 5800X.

Your comparison shouldn't be run. Period. It has different OSes. that's like comparing a plum to a hamster. it doesn't work the same way.

 

[EatingDirt]

9% is a negligible performance improvement and literally very disappointing, in the ballpark of simple rebrands.
No user would ever notice better user experience with this.

Why does AMD even waste its time and resources, while instead doesn't pull the next generation Zen 4 CPUs launch forward?

9% is a semi-generational jump in terms of CPU's(See intel & Zen to Zen+), except this is just the same CPU with added cache. This CPU can act as the final upgrade for gamers on the AM4 platform, and is a significant upgrade in gaming for anyone on Zen, Zen+ or Zen2 CPU's.

Asking "Why not just Zen 4" shows ignorance on how fabrication-reliant technology works. First off, this CPU is not made on the same process as Zen 4, so it doesn't cut into Zen 4 production at all. Other reasons Zen 4 can't be pushed forward: Lack of TSMC capacity, Adequate supply of chipsets for motherboards, Adequate supply of DDR5. There's more, but I hope you get the picture.

 

[Valantar]

9% is a semi-generational jump in terms of CPU's(See intel & Zen to Zen+), except this is just the same CPU with added cache. This CPU can act as the final upgrade for gamers on the AM4 platform, and is a significant upgrade in gaming for anyone on Zen, Zen+ or Zen2 CPU's.

Asking "Why not just Zen 4" shows ignorance on how fabrication-reliant technology works. First off, this CPU is not made on the same process as Zen 4, so it doesn't cut into Zen 4 production at all. Other reasons Zen 4 can't be pushed forward: Lack of TSMC capacity, Adequate supply of chipsets for motherboards, Adequate supply of DDR5. There's more, but I hope you get the picture.

I was about to bring up the same point: that "bringing forward Zen4" isn't feasible in a strictly supply-constrained market. It's not like AMD can ask TSMC to push their 5nm production forward by a few months - they have zero free capacity, with everything being booked for likely a year if not more. I expect Zen4 chiplets to be in mass production now, but that's the "building up stock before launch" stage, and that doesn't mean they have packaging, a new IOD, or all the other things necessary for Zen4 to launch ready. Let alone AM5 motherboards, as there's no way to get DDR5-only Zen4 onto AM4. The 5800X3D is a last hurrah for the AM4 platform, and AMD testing out a new, exotic production and packaging technique, that might also bring tangible benefits for gamers looking for high framerates. If it doesn't bring much else to the table, that's not only fine, but expected.

 

[SL2]

Before going into childish personal attacks, noone ever looks at synthetic multicore results and expects gaming results from them. For all the duration of multicore processors, 17 years. Everyone reads my comment about Adler Lake scores as single core for gaming and multi core for productivity.

That's the kind of replies you can expect when you make annoying replies like you did. I deliberately picked some data that no one believes reflects gaming performance, and you choose to react childish to it, as if I was clueless about it despite that I clearly just pointed this out.

Or, if I would use your vocabulary:
"And why, for God's sake, would you think that GB would reflect the performance of X3D as good as it does on Alder?"
Yeah, a bit childish. I'd say we're eve now.

I stated that 9% uplift in multicore AND REGRESSION in single core Geekbench result is a very bad prognosis for gaming increase that AMD is promising.

There's no reason to believe that the X3D would be slower in gaming, it's not very likely. (Well and neither is 20 % faster or whatever AMD said)
If you think you can make such a prognosis for X3D out of GB I'd say you're in a minority here. Just because GB was right about Alder it doesn't mean that GB is right every time, and it's especially not expected in such an odd product like X3D.

Now if Raptor Lake, or Raphael, would have a lower single GB score, that would surprise me.

It's not just me. In case you haven't read the source:

The single-core performance is almost the same but 3D V-Cache gains would mostly be apparent in cache-dependent workloads such as games rather than core-performance optimized benchmarks such as Geekbench.

The hardware improvements of X3D are nothing like the improvements of Alder. Maybe we should compare it to Broadwell S instead? Seems more fitting to me, even though it wasn't extra L3.
It was about the same as the 4790K in both GB tests but always faster in games (except for C6) with an average of 19 % in low settings and resolutions, even though the latter runs up to 700 MHz faster.

In other words, neither single GB 4 or 5 could predict gaming performance from the extra cache back then.

In the end I blame Gbench, it's the same story every time. It should be banned from unreleased CPU's lol...

9% is a negligible performance improvement and literally very disappointing, in the ballpark of simple rebrands.
No user would ever notice better user experience with this.

Why does AMD even waste its time and resources, while instead doesn't pull the next generation Zen 4 CPUs launch forward?

Why don't you just read some background info instead of chasing higher model numbers? We've heard your Zen4 chanting for ages now. Oh and give me a sad face while you're at it.

 

[TechLurker]

This also seems to show how much more refined the newer 5800X3D bins are over older 5800X bins. I wonder if AMD is also applying newer steppings to more recent production batches, and if said batches are also similarly improved over earlier steppings.

 

[Ruru]

This also seems to show how much more refined the newer 5800X3D bins are over older 5800X bins. I wonder if AMD is also applying newer steppings to more recent production batches, and if said batches are also similarly improved over earlier steppings.

Wouldn't mind if the upcoming 5700Xs are also more refined than current 5800Xs.. as that's where I'm going for, though 5800X3D is definitely the most interesting one.

 

[Jism]

So much cache but little uplift.. shows l2 cache is more important thats why intel will increase it for raptor lake

You know that the CPU's clocks where lowered with a good 200Mhz right?

9% "Faster" while having lower speed clocks, shows there's plenty of oppertunity in these chips.

 

[ThrashZone]

You know that the CPU's clocks where lowered with a good 200Mhz right?

9% "Faster" while having lower speed clocks, shows there's plenty of oppertunity in these chips.

Hi,
That might be a down fall to seeing this chip can not be oc'ed so default is all she can do good or bad.

 

[ARF]

9% is a semi-generational jump in terms of CPU's(See intel & Zen to Zen+), except this is just the same CPU with added cache. This CPU can act as the final upgrade for gamers on the AM4 platform, and is a significant upgrade in gaming for anyone on Zen, Zen+ or Zen2 CPU's.

Asking "Why not just Zen 4" shows ignorance on how fabrication-reliant technology works. First off, this CPU is not made on the same process as Zen 4, so it doesn't cut into Zen 4 production at all. Other reasons Zen 4 can't be pushed forward: Lack of TSMC capacity, Adequate supply of chipsets for motherboards, Adequate supply of DDR5. There's more, but I hope you get the picture.

Ignorance as a word used in this reply is very hostile action in itself. And when you ask yourself why I won't engage in replying to your trolling, you must know now..

 

[Jism]

Hi,
That might be a down fall to seeing this chip can not be oc'ed so default is all she can do good or bad.

They dont have a seperate voltage rail for the Cache on top; so it's hooked to the CPU VCore. Now that cache cant handle things above 1.35 so it seems so they completely disabled any OC features which makes perfect sense. If your cache is toast the rest of your chip is, because it coud'nt produce proper results anymore.

But, i think as long as you have a board with a external BCLK generator you still should be able to increase max boost clocks.

 

[ThrashZone]

They dont have a seperate voltage rail for the Cache on top; so it's hooked to the CPU VCore. Now that cache cant handle things above 1.35 so it seems so they completely disabled any OC features which makes perfect sense. If your cache is toast the rest of your chip is, because it coud'nt produce proper results anymore.

But, i think as long as you have a board with a external BCLK generator you still should be able to increase max boost clocks.

Hi,
Some higher end board manufactures may give more options for sure

Also uncleweb for throttlestop too

 

[Aquinus]

Okay people. I understand the whole "GeekBench is garbage," argument, but it's an okay thing to use when you're comparing CPUs of the same architecture because you're comparing apples to apples. What we're seeing with these numbers is improved efficiency with sufficient memory pressure (both in terms of how much is used and/or how often it's accessed.) Once again, this is an improvement without increasing clocks, which means that this chip is more efficient than its predecessor. It goes to show that with how memory hungry applications can be these days, that the gain from a larger cache has far more tangible benefits than we would have otherwise thought.

Like I said, go check out that EPYC review on Phoronix. In the server space, this kind of thing is making improvements far larger than a 9% gain. It does make me wonder how well more cache would scale.

 

[ThrashZone]

Hi,
Thought it was user benchmark sucks, this one geek bench is just no where on any benchmark forum except this "rumor" news thread

 

[Chrispy_]

I hate that Geekbench is always the first leaked performance metric.
It's such a useless benchmark!

 

[Valantar]

I hate that Geekbench is always the first leaked performance metric.
It's such a useless benchmark!

The correlation between "easiest to set up and run" and "least useful" is quite strong.

 

[maxfly]

If this person actually has one, why wouldn't they run an entire battery of benchmarks and tests?

 

[Why_Me]

This cpu looks like it has the makings of being a good upgrade for those that already own an AMD board.

 

[Minus Infinity]

Again way too late to be of interest to me. No matter what improvements 5800X3D shows, the 7800X will blow it away and is due Q3. Now if this were released 12 months ago, I may have updated my 3700X to one.

 

[Chrispy_]

Again way too late to be of interest to me. No matter what improvements 5800X3D shows, the 7800X will blow it away and is due Q3. Now if this were released 12 months ago, I may have updated my 3700X to one.

I thought AMD were delaying Zen4 because DDR5 is suffering production teething troubles and there's a massive supply shortage - one so bad that Alder Lake would be dead in the water without DDR4 boards.

The 7800X may kick ass when it arrives, but latest info says it's been pushed back to holiday season, so end of Q4. If you can wait another 9 months then you're fine, but until then the 5800X provides some competition to Alder Lake.

 

[ThrashZone]

Again way too late to be of interest to me. No matter what improvements 5800X3D shows, the 7800X will blow it away and is due Q3. Now if this were released 12 months ago, I may have updated my 3700X to one.

Hi,
In gaming ?
You might be dreaming