News AMD Ryzen 9000 CPU family compared in Cinebench — purported scores for the 9900X, 9700X, and 9600X shared

[Admin]

Cinebench results of AMD's Ryzen 9900X, 9700X, and 9600X get published ahead of reviews.

AMD Ryzen 9000 CPU family compared in Cinebench — purported scores for the 9900X, 9700X, and 9600X shared : Read more

 

[TheHerald]

I wish I had bet money on this! 9600x around 12600k performance....

 

[bit_user]

IMO, the single-threaded improvement is most interesting, especially given the lack of clockspeed increases.

As for the multi-threaded cases, I wonder if it's being held back by the speed of the DDR5 memory being used. The MT case should benefit from the compound effect of Zen 5's IPC improvements and the efficiency improvements touted by AMD. Therefore, I'd expect it to improve by more than the single-threaded case.

 

[bit_user]

I wish I had bet money on this! 9600x around 12600k performance....

BTW, the R5 9600X is rated at 65 W, while the i5-12600K is rated at 125W.

Also, what's your source for an authoritative score for the i5-12600K? In this WCCFTech article, they list a score of 15982, which 17037 beats by 6.6%.

AMD Ryzen 7 9700X & Ryzen 5 9600X "Zen 5" CPUs Tested At Both Default & PBO Configs, 2% Uplift in Cyberpunk

AMD's Ryzen 7 9700X & Ryzen 5 9600X "Zen 5" CPUs have been tested in newly leaked benchmarks at both default & PBO configurations.

wccftech.com

On single-threaded CineBench R23 performance, the above WCCFTech article shows the R5 9600X being 4.0% faster than even the i5-14600K (they don't list a ST score for the i5-12600K). If we take the ST score from the Toms article, the gap jumps to 8.0%!

 

[TheHerald]

BTW, the R5 9600X is rated at 65 W, while the i5-12600K is rated at 125W.

And that is relevant to the performance, how?

BTW, the R5 9600X is rated at 65 W, while the i5-12600K is rated at 125W.

Also, what's your source for an authoritative score for the i5-12600K? In this WCCFTech article, they list a score of 15982, which 17037 beats by 6.6%

Both TPU and tomshardware.com have the 12600k at 17 to 18k. The 16k is definitely false, but let's say it's correct. Still a 6.6% difference is what I call minor, no?

The 17k you are using btw is with PBO.

 

[bit_user]

And that is relevant to the performance, how?

It's a multi-threaded benchmark, which is going to be power-limited on the i5-12600K.

The 17k you are using btw is with PBO.

Oh, good catch. Yes, Toms just took the PBO score. That explains the discrepancy between the two articles.

So, I guess we'd have to know how much power it's using in PBO mode, or just stick with the WCCFTech "(default)" scores.

 

[TheHerald]

It's a multi-threaded benchmark, which is going to be power-limited on the i5-12600K.


Oh, good catch. Yes, Toms just took the PBO score. That explains the discrepancy between the two articles.

So, I guess we'd have to know how much power it's using in PBO mode, or just stick with the WCCFTech "(default)" scores.

According to TPU the 600k scores 17500 while limited to 125w.

 

[Reality_checker]

Not final, but also unlikely to get a lot better.

The 9950X is the only interesting SKU anyway.

 

[-Fran-]

I wish they reported the clocks and power draw while testing CB. Well, in the leak at least...

Not too much longer for getting official numbers, so all points to this being a decent enough uplift with a lot of bragging rights on the efficiency department (I think).

Regards.

 

[vinay2070]

That is less than what I expected TBH. Thats around 10% on average. So the MT on 9950x will be around 7% then?

 

[bit_user]

That is less than what I expected TBH. Thats around 10% on average. So the MT on 9950x will be around 7% then?

As a matter of fact, if you look at this chart, it has a bar specifically for Cinebench R23 which claims 17% IPC improvement.

​

Therefore, such a small real world improvement suggests these CPUs are running at substantially lower clocks than their predecessors, which seems weird, given the supposed efficiency improvements.

 

[YouFakEyeWhatch]

And that is relevant to the performance, how?



Both TPU and tomshardware.com have the 12600k at 17 to 18k. The 16k is definitely false, but let's say it's correct. Still a 6.6% difference is what I call minor, no?

The 17k you are using btw is with PBO.

Actually the 12600k with a mild oc (5.4GHz P-Core) scores 21500MT/2100ST. In comparison to last Gen's 7600x with a mild oc(5.6GHz A/C) that scores 16500MT/2050ST. The 9600x chip will not beat intel's old 12600k in multi-threaded even with overclock unless you go below ambient for a real overclock. Either way at the end of the day 9000 series is not even on part with 13/14th gen, it looks like it may have better IPC though, but unfortunately Ryzen is still not hitting 6GHz above ambient this generation. Even with the "perhaps" improved IPC on Ryzen 9000, scaling is poor once you go below ambient.

 

[vinay2070]

As a matter of fact, if you look at this chart, it has a bar specifically for Cinebench R23 which claims 17% IPC improvement.

​

​

​

Therefore, such a small real world improvement suggests these CPUs are running at substantially lower clocks than their predecessors, which seems weird, given the supposed efficiency improvements.

I hope so you are right.

 

[mac_angel]

I was expecting a lot higher. Those scores don't seem that impressive compared to the 13900K and the 14900K, which have another 10K in points above the 9900X

 

[KyaraM]

IMO, the single-threaded improvement is most interesting, especially given the lack of clockspeed increases.

As for the multi-threaded cases, I wonder if it's being held back by the speed of the DDR5 memory being used. The MT case should benefit from the compound effect of Zen 5's IPC improvements and the efficiency improvements touted by AMD. Therefore, I'd expect it to improve by more than the single-threaded case.

According to what the article lists, they seem between 400 to 800 MHz slower compared to their direct respective predecessor, which I find very weird... heck, even my laptop 7640HS clocks higher than the 9900X. Though the article doesn't say if the listed speeds are actual clocks during the test, max rated, base, PBO, etc. Though it doesn't much matter provided that the old chips use the same metrics.

Actually the 12600k with a mild oc (5.4GHz P-Core) scores 21500MT/2100ST. In comparison to last Gen's 7600x with a mild oc(5.6GHz A/C) that scores 16500MT/2050ST. The 9600x chip will not beat intel's old 12600k in multi-threaded even with overclock unless you go below ambient for a real overclock. Either way at the end of the day 9000 series is not even on part with 13/14th gen, it looks like it may have better IPC though, but unfortunately Ryzen is still not hitting 6GHz above ambient this generation. Even with the "perhaps" improved IPC on Ryzen 9000, scaling is poor once you go below ambient.

I'm not sure if I would call 500 MHz above max ST boost "mild". No, scrap that, that's definitely NOT a mild OC. Heck, I only OC my 12700K to 4.9GHz/4.0GHz allcore... which is below max ST speed. THAT is a mild OC.

 

[bit_user]

I was expecting a lot higher. Those scores don't seem that impressive compared to the 13900K and the 14900K, which have another 10K in points above the 9900X

The 9950X is the flagship model, though. With the 9900X, you're putting 12/24 cores/threads up against 24/32. Obviously, that's going to be a big mountain to overcome. The 9950X matches them in threads (32), but still with just 16 cores.

 

[TheHerald]

Actually the 12600k with a mild oc (5.4GHz P-Core) scores 21500MT/2100ST. In comparison to last Gen's 7600x with a mild oc(5.6GHz A/C) that scores 16500MT/2050ST. The 9600x chip will not beat intel's old 12600k in multi-threaded even with overclock unless you go below ambient for a real overclock. Either way at the end of the day 9000 series is not even on part with 13/14th gen, it looks like it may have better IPC though, but unfortunately Ryzen is still not hitting 6GHz above ambient this generation. Even with the "perhaps" improved IPC on Ryzen 9000, scaling is poor once you go below ambient.

That's definitely not a mild oc and you are most likely hitting rapid degradation voltages for those clocks. It's not a realistic score at all, it's just for benchmarking. You aren't going to use the cpu for any serious workload running at those settings cause it will require an RMA within the day.

 

[TheHerald]

The 9950X is the flagship model, though. With the 9900X, you're putting 12/24 cores/threads up against 24/32. Obviously, that's going to be a big mountain to overcome. The 9950X matches them in threads (32), but still with just 16 cores.

If it price matches a 24/32 cpu then it should be beating or at least being close. If the reason it's not close is the lack of cores then there should be a lack of $ from the asking price reflecting that.

 

[Dementoss]

"Purported scores"?

I'll wait until we get some verifiable test results, before making a judgement.

 

[Guardians Bane]

If it price matches a 24/32 cpu then it should be beating or at least being close. If the reason it's not close is the lack of cores then there should be a lack of $ from the asking price reflecting that.

Ok then, if you pit Intel 14900k against the 9900x, then I guess that means Intel has nothing higher? Either way, the 9950x is the flagship. Your price thing? Eh, to a degree. But the other way to look at it is 9950x is flagship (expensive yes) and Intel 14900k is flagship , competes with 9900x then the top flagship sku from AMD has no competition. Your price thing just gave AMD a top sku with no direct competition.

All we gotta do is wait for real world test results and reviews. It won't hurt to see the real thing. Then we can go back and forth and say who gets crowned. I don't care either way. I'm waiting to build my new PC after Intel's next get drops so I can compare Intel vs AMD. I will not buy 13th or 14th gen at all. I'm patient and can wait though.

 

[TheHerald]

Ok then, if you pit Intel 14900k against the 9900x, then I guess that means Intel has nothing higher? Either way, the 9950x is the flagship. Your price thing? Eh, to a degree. But the other way to look at it is 9950x is flagship (expensive yes) and Intel 14900k is flagship , competes with 9900x then the top flagship sku from AMD has no competition. Your price thing just gave AMD a top sku with no direct competition.

All we gotta do is wait for real world test results and reviews. It won't hurt to see the real thing. Then we can go back and forth and say who gets crowned. I don't care either way. I'm waiting to build my new PC after Intel's next get drops so I can compare Intel vs AMD. I will not buy 13th or 14th gen at all. I'm patient and can wait though.

Well going by 13th gen vs zen 4, the 7900x had an msrp of 549 vs 599 for the 13900k,so yes, the 900k competed with the 900x.

Yes, intel doesnt have a cpu as expensive as the msrp of the 950x, unless you count the special ks editions.

 

[bit_user]

Well going by 13th gen vs zen 4, the 7900x had an msrp of 549 vs 599 for the 13900k,so yes, the 900k competed with the 900x.

MSRP is a fiction, especially years after launch. The only thing that matters is street price.

The rumored street price for the 9950X is $500, in which case it's very much a matchup for the i9-14900K!

Let's wait for the official word on pricing.

 

[TheHerald]

MSRP is a fiction, especially years after launch. The only thing that matters is street price.

The rumored street price for the 9950X is $500, in which case it's very much a matchup for the i9-14900K!

Let's wait for the official word on pricing.

MSRP is what the company thinks their cpu is worth / what it competes against. Street price is what the customer thinks said cpu is worth.

In any case I don't think the 9950x will be at 500, not any time soon. If it is indeed 500 that's awesome, it will be faster in MT, a chunk more efficient and probably equal in gaming and ST performance. Plus - hopefully - none of the recent issues that plague intel. Would be a no brainer at that price.

 

[bit_user]

MSRP is what the company thinks their cpu is worth / what it competes against.

To the extent this is true, it reflects the moment in time when the product is launched. Comparing the MSRP of a product launched in 2022 with that of a product launched more than a year later doesn't make a whole lot of sense. The earlier product's MSRP definitely wasn't set with the expectation that it represented its market value so far into the future.

Street price is what the customer thinks said cpu is worth.

Not really. Street prices are ultimately limited by what the manufacturer is willing to sell it for. They're continually weighing estimated increased sales volume from price drops against the loss of margin. The market isn't necessarily as responsive to price drops as the manufacturer's bottom line is to loss of margin. Let's not pretend these products don't have real costs.

In any case I don't think the 9950x will be at 500, not any time soon.

True. The breaking news is that it's launching at $600 (MSRP), but the R9 9900X is launching at $450. So, that puts the i9-14900K in between (but decidedly closer to the R9 9950X, unless Intel makes further price cuts).

 

[TheHerald]

To the extent this is true, it reflects the moment in time when the product is launched. Comparing the MSRP of a product launched in 2022 with that of a product launched more than a year later doesn't make a whole lot of sense. The earlier product's MSRP definitely wasn't set with the expectation that it represented its market value so far into the future.

Man the 13900k is a replica of the 14900k and wasn't released a year later. Come on now...

Not really. Street prices are ultimately limited by what the manufacturer is willing to sell it for. They're continually weighing estimated increased sales volume from price drops against the loss of margin. The market isn't necessarily as responsive to price drops as the manufacturer's bottom line is to loss of margin. Let's not pretend these products don't have real costs.

But the opposite isn't true. A company isn't lowering the price of its products if they are selling at or above expectations. Zen 4 dropped like a rock because apparently it wasn't selling as expected, so ignoring its msrp and just comparing street prices is kinda meh.