Question What if the real reason for both the lack of availability as well as the lower-latency of the 3300X is because...

[VirtualLarry]

They are dud Zen3 dies!

This could also explain the crazy memory OCs on the 3300X that some people found.

 

[DrMrLordX]

Not likely. If they are, Zen3 is going to be a real pig. Zen3 is supposed to have +15% IPC across the board. 3300x doesn't show any increase in IPC.

 

[NostaSeronx]

Would be noticed like the A8-7600: https://valid.x86.fr/dlvnfd <-> A8-7680: http://valid.x86.fr/yvnrw0

 

[VirtualLarry]

If they are, Zen3 is going to be a real pig.

Well, Nosta (bless his heart) did recently post about a split in Zen3 development, and a "Fake" Zen3, with "a few changed blocks", and a "Real" Zen3, that was the new architecture.

I know what Nosta says, rarely comes true, but what if that were true? That the "Zen3" that is going to show up soon, or by the end of 2020, is going to be the "Fake" Zen3, with only a few changed blocks from Zen2, and the 3300X are rejected "Fake Zen3" dies, that AMD decided to sell.

Has anyone compared IPC at ISO clocks and or power limits for the 3300X to Zen2 dies? They are so rare, hard to find to get a hold of to make a comparison.

I know, I know, a bit conspiratorial, but why else is the availability of the 3300X so low? And yet, the 3100 is available. Seems slightly suspicious, when taken with the "other" factors that make the 3300X seem uniquely faster than all other Zen2 CPUs, for gaming IPC.

 

[DrMrLordX]

If AMD only sells Zen3 with "a few changed blocks" then there will be outrage. And I wouldn't lend that rumour any credence.

 

[NostaSeronx]

Even if it is fake "Zen3" it would have:

Fast Short REP MOV, PCID/INVPCID, 256-bit AES/CLMUL. MCA/EDAC changes. Individual per-core voltage and frequency settings.

It would be noticed if 3300X/3100 had those and 19.20 in CPU-Z.

Basically the major differences between:
Genesis = 00h-0Fh => SP3 (Still, Milan)
Stones = 10h-1Fh => SP5 (Was, Genoa) // 1st gen chiplet - 2017 => 2nd gen chiplet - 2019 => 3rd gen chiplet - 2021(Stones/XYZ)/2022(Floyd/Raphael)
Vermeer = 20h-2Fh => AM4
Is that they share the processor_core block with Family 17h 30h-3Fh and 70h-7Fh.
Basically, a X3D-capable Improved Zen2 processor.

While, 30h and greater uses a new processor_core block. Plus more thermal-aware core design that can still get >3 GHz in stack.

XYZ = any one that isn't Raphael: Michelangelo, Raphael and Leonardo da Vinci

 

[A///]

Do you guys ever read before posting? There's been countless benches done since the product released. It's a high demand product. There is a pandemic. The 3300X matches or just barely beats the 7700K in a myriad of benchmarks. It just about edges out the 7700K in Geek Bench.

AMD Ryzen 3 3300X beats Intel Core i7-7700K in Geekbench - KitGuru

AMD recently announced the new 4C/8T Ryzen 3 processors. Although the CPUs haven't launched yet, the

www.kitguru.net

AMD Ryzen 3 3300X vs. Intel Core i7-7700K - Benchmark

AMD Ryzen 3 3300X vs. Intel Core i7-7700K - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

R20 Single Core

3300X: 495
7700K: 466

R20 Multi-Core

3300X: 2588
7700K: 2257

R15 Single Core:

3300X: 200
7700K: 200

R15 Multi-Core

3300X: 1129
7700K: 995


Compared to the 10th gen i3-10320:

AMD Ryzen 3 3300X vs. Intel Core i3-10320 - Benchmark

AMD Ryzen 3 3300X vs. Intel Core i3-10320 - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

R20 Single Core

3300X: 495
10320: 477



R20 Multi-Core

3300X: 2588
10320: 2444

R15 Single Core:

3300X: 200
10320: 206

R15 Multi-Core

3300X: 1129
10320: 1080



Zen cores were designed with multicore efficiency in mind. There should be a noticeable gap in multicore performance between a 3300X vs a 3600X, for example. In other words, the 3300X should out perform a prior gen 6 core part going by a 10-15% increase in IPC.

AMD Ryzen 5 3600X vs. AMD Ryzen 3 3300X - Benchmark

AMD Ryzen 5 3600X vs. AMD Ryzen 3 3300X - Benchmark, Geekbench 5, Cinebench R20, Cinebench R23, Cinebench R15 and FP32 iGPU (GFLOPS) benchmark results

www.cpu-monkey.com

R20 Single Core

3300X: 495
3600X: 501

R20 Multi-Core

3300X: 2588
3600X: 3751

R15 Single Core:

3300X: 200
3600X: 204

R15 Multi-Core

3300X: 1129
3600X: 1647



Compared to the 3100, which uses two CCDs, there's anywhere from a few points to less than 220 pts. difference in CB 15 and 20 tests. If the 3100 uses dud Zen 2, and the 3300X uses dud Zen 3, meaning a single 8 core chiplet with 4 terrible cores disabled, then AMD has failed to improve much on Zen 2 despite fixing cache, latency due to cache, a redesigned core, and terrible clock regressions over Zen 2. If that held water, AMD's stock would plummet by at least 30-50%. Their datacenter wouldn't be much better, either.

The 3300X literally performs better in general tasks and games. You're wondering why the 3100 is always available vs the 3300X. Why the heck were AMD Bulldozers available and Core processors almost always sold out each big release?

If they're Zen 3 duds, then not only has AMD massively failed to deliver a noticeable IPC increase, but they also inherited a clock regression. You're implying that not only is the 3300X built on dud Zen 3 chiplets, but that also the 3100 would be, too. As both the 3100 and 3300X have less cores than the 3600X, they should in theory be capable of boosting higher than the 3600X.

Steve over at Gamers Nexus did not only a lengthy bench article but also a video where he highly recommended the 3300X over the 3100. In most games, the 3300X barely has more than a few frames more than the 7700X. You're suggesting that with all the fancy updates, and 4 generations, Zen still can't put a nice gap between them and Intel's then 4/8 processor from almost 4 years ago.

AMD Ryzen 3 3300X CPU Review vs. 3100 Benchmarks: An R3 is Enough for Gaming & Clock-for-Clock

This is the big one: In this review, we’re benchmarking the AMD R3 3300X $120 CPU, but we’re specifically interested in the real-world impact of the CCX-to-CCX communication latency in the Ryzen 3 3100 versus the Ryzen 3 3300X at the same overclocked frequency of 4.4GHz. -

www.gamersnexus.net

What's next? The XT's were dud Zen 4 engineering sample chiplets?

 

[Arkaign]

AMD really dropped the ball with 3300X. Don't get me wrong, it's a fine product, but availability is so poor it's basically vaporware, to the point Amazon no longer even has a page for it outside of 3rd party scalpers (whom also don't have any lol). My Dallas Micro Center hasn't had any in months, since release in fact. And the reports of potential future availability puts it squarely in Zen3 release window lol. So, it will be last gen by the time they (maybe) get around to getting some available.

Given the widely available 3600/3700/3800/X/XT stuff, it seems they prioritized more profitable SKUs.

 

[Blue_Max]

The 3300X benches almost exactly the same as the Ryzen 3600 in most games. Like, the vast majority of them! The 3100 trails way behind in comparison. For gaming, it's just an incredible value... I really should've grabbed one when my local shop got a small batch in last month!

 

[blckgrffn]

The 3300X benches almost exactly the same as the Ryzen 3600 in most games. Like, the vast majority of them! The 3100 trails way behind in comparison. For gaming, it's just an incredible value... I really should've grabbed one when my local shop got a small batch in last month!

When a 3600 has been reliably $20-$30 more in the same time frame? They always seemed like solid business PC CPUs to me but a step too far down the ladder for enthusiasts.

And with the much more broadly available 3400G looking over its shoulder for the absolutely lowest budget possible gaming crowd, it seemed even less useful. Especially given that even used GPUs that surpass the onboard GPU of the 3400G cost easily over $50 ¯\_(ツ)_/¯

Clearly wasn't a CPU aimed at me

 

[chrisjames61]

Well, Nosta (bless his heart) did recently post about a split in Zen3 development, and a "Fake" Zen3, with "a few changed blocks", and a "Real" Zen3, that was the new architecture.

I know what Nosta says, rarely comes true, but what if that were true? That the "Zen3" that is going to show up soon, or by the end of 2020, is going to be the "Fake" Zen3, with only a few changed blocks from Zen2, and the 3300X are rejected "Fake Zen3" dies, that AMD decided to sell.

Has anyone compared IPC at ISO clocks and or power limits for the 3300X to Zen2 dies? They are so rare, hard to find to get a hold of to make a comparison.

I know, I know, a bit conspiratorial, but why else is the availability of the 3300X so low? And yet, the 3100 is available. Seems slightly suspicious, when taken with the "other" factors that make the 3300X seem uniquely faster than all other Zen2 CPUs, for gaming IPC.

What Nostra says never comes true. Larry, seems like you either like conspiracy theories or want AMD to fail.

 

[Blue_Max]

When a 3600 has been reliably $20-$30 more in the same time frame? They always seemed like solid business PC CPUs to me but a step too far down the ladder for enthusiasts.

Around here it's a $100 difference. $175 vs $275 Canadian. Considering the gaming performance (my preferred activity) is almost identical, there's a good reason budget gamers snap them up the moment they hit the shelves.

 

[blckgrffn]

Around here it's a $100 difference. $175 vs $275 Canadian. Considering the gaming performance (my preferred activity) is almost identical, there's a good reason budget gamers snap them up the moment they hit the shelves.

Well, it's back to the same here but in the good old days right after launch that wasn't the case at all. I don't know that any real number of them hit the channel since then here. The 3600 pricing situation only changed in the last couple of weeks, presumably as we are getting so close to the Zen 3 drop. It was pretty reliably around $150 at MC with a $20 board credit and is now $180 with a $20 combo credit. Still a fantastic deal given the relative performance IMO.

Hopefully it's a drop and not a paper launch