Discussion AMD Ryzen MegaThread! FAQ and Resources

[juanrga]

The 6950X has a turbo is lower than the base clock of the 1800X.

Zen @3.6GHz ~ Haswell-E @3GHz ~ Broadwell-E @3GHz

Haswell-E and Broadwell-E can clock higher than 3.5GHz through Turbo Boost 2.0 and 3.0 respectively.

Turbo 3 is not being used here and the reason why the 6850K beats the 6900K and the 6800K beats the 6950X.

Why would they disable only one aspect of turbo core? Please link where they clearly state they disabled any of the turbo aspects of any chip.

Turbo 3 requires a special board to work and a driver to work, and it has to be enabled and configured.

The proof that Turbo 3 is not enabled is evident from the graph data, as Gon Freecss already admitted here

 

[juanrga]

Ryzen is a different architecture at times when i compared the 1500X to 7700K i saw more IPC with ryzen but this was just as often as when i saw way less IPC(like 20-40%). On average who's to say Amd claims 10% within kabylake which is basically Haswell i say it depends, on average probably around Haswell maybe a little less.

I said it a LONG time ago but even if Ryzen got Sandy-ivy level IPC i'd call it a success that stands true today.

Average IPC about 10% behind Broadwell on compute ( 259.5 / 235.8 = 1.10 )

Average IPC about 20% behind Broadwell on games ( 209 / 173.3 = 1.21 )

10--20% behind Broadwell implies the IPC is well below Haswell. This is the reason why RyZen has to be clocked higher to match Haswell

And it is the reason why the R5 1400 and 1500X perform like Sandy and Ivy Bridge i7s more or less.

On compute RyZen IPC is a ~13% behind Skylake/Kabylake:

As you would expect, AMD still lags in IPC to Intel, so a 4.0 GHz AMD chip can somewhat compete in single threaded tests when the Intel CPU is around 3.5-3.6 GHz, and the single thread web tests/Cinebench results show that.

http://www.anandtech.com/show/11244/the-amd-ryzen-5-1600x-vs-core-i5-review-twelve-threads-vs-four/17

4 / 3.6 = 1.11
4 / 3.5 = 1.14

RyZen Compute IPC: ~6% above Sandy, ~10% below Broadwell, ~13% below Skylake.
RyZen gaming IPC: worse than that.

 

[XBloodyR]

@juanrga

Those are pretty good results considering its just the single threaded performance when ryzen is all about the multi-cores not to mention we dont really know anything about the benchmarking systems they had, they could have had 2400 mhz ram as far as we know and the older bios.
Another thing they just seem to ignore is that the higher overclocked kaby lakes like on 4.8-5 Ghz do not seem to benefit as much from it anymore because they are at the limits of their architecture as you see by intels gradually non existent improvements in new generations. I would not be willing to spent 420-450€ for the cpu+cooler just for those 30% speed kaby lake has over my overclocked sandy bridge.

 

[juanrga]

@juanrga

Those are pretty good results considering its just the single threaded performance when ryzen is all about the multi-cores not to mention we dont really know anything about the benchmarking systems they had, they could have had 2400 mhz ram as far as we know and the older bios.

The above graphs aren't for single-thread, but for well-threaded benches; that is the reason why the 10C Broadwell got the maximum score in compute, above both 8C Broadwell and 8C RyZen. They tested different RAM configurations, BIOS, and more. Performance change from some few percents to nothing as I demonstrated in former posts. I will repeat the graph with the performance improvement that brings the new BIOS/AGESA

Another thing they just seem to ignore is that the higher overclocked kaby lakes like on 4.8-5 Ghz do not seem to benefit as much from it anymore because they are at the limits of their architecture as you see by intels gradually non existent improvements in new generations. I would not be willing to spent 420-450€ for the cpu+cooler just for those 30% speed kaby lake has over my overclocked sandy bridge.

Intel did hit a performance wall. The evolution of x86 has stopped (*) and AMD hits the same wall and Zen+ (Zen2) is will bring minor IPC gains over Zen. I guess between 5% and 10%.

(*) The only evolution what remains are the non-x86 ISAs like AVX and "moar cores".

 

[-Fran-]

Mixing different data points leads to wrong conclusions.

Cheers!

 

[XBloodyR]

What i meant by single threaded is the following.

Quote:
As you would expect, AMD still lags in IPC to Intel, so a 4.0 GHz AMD chip can somewhat compete in single threaded tests when the Intel CPU is around 3.5-3.6 GHz, and the single thread web tests/Cinebench results show that.


http://www.anandtech.com/show/11244/the-amd-ryzen-5-160...

4 / 3.6 = 1.11
4 / 3.5 = 1.14

RyZen Compute IPC: ~6% above Sandy, ~10% below Broadwell, ~13% below Skylake.
RyZen gaming IPC: worse than that.

 

[salgado18]

Allright, this "IPC is lower than / higher than" discussion is not taking anyone anywhere, especially because each one can see what they want to see.

My point of view? IPC is not as good as Intel:

But, like the music says, "I don't care, I love you (Ryzen)":

The world is not single-threaded anymore! IPC is only a part of the story, anything that needs CPU performance is doing its best to go threaded. In that territory, Ryzen is king.

And in that sense, AMD made an awesome product.

Want better IPC? Wait for Zen+. Intel may have reached a wall (I still think they were holding back, though), but AMD hasn't. It's a new CPU, with lots of room for improvement.

So, please, can we stop this months-old argument of IPC, especially when there are so few benchmarks testing this on the web? And no, one single bench on one single program tells nothing:

Edit: I mean, we are here to discuss Ryzen, but we are going in circles, always using the same couple of single-threaded tests around the web to justify our points of view. Until we see a reviewer or user do various single-threaded benchmarks, and compare to the same tests on Intels at the same clocks, we can't go any further, and any conclusion will be speculation.

 

[8350rocks]

Mixing different data points leads to wrong conclusions.

Cheers!

/portion of the thread saying 10% behind broadwell

 

[Gon Freecss]

@juanrga;

The 1500X is actually better than the 2600K when it comes to gaming, especially the minimums, when both are at stock.

Ryzen is between Sandy and Ivy in gaming, and between Ivy and Haswell when it comes to professional work.

 

[jdwii]

@juanrga;

The 1500X is actually better than the 2600K when it comes to gaming, especially the minimums, when both are at stock.

Ryzen is between Sandy and Ivy in gaming, and between Ivy and Haswell when it comes to professional work.

Now that i agree with 100%. I've been looking at as many benchmarks as possible and one can not deny that Ryzen isn't behind Haswell in gaming in most titles. Its actually frustrating to me as i want to switch but i don't want to get less FPS in my open world titles fallout 4 for example gets 20% less FPS on ryzen. I know its a crappy coded game but that doesn't mean i want less performance.

I also was able to get my chip at 4.7Ghz where it looks like Ryzen will probably be 3.8-4.1Ghz. In my case i'll be getting less IPC then haswell in games as well as 15% less frequency. However i can get twice the cores but games on ryzen just don't scale the same as they do with a 6900K least the titles i'm looking at.

IMO Ryzen or kabylake look better for gamers with sandy or older based platforms. I can't justify upgrading to a 7700K that's for sure, if i switch to ryzen it will be for fun more for a logical reason.

 

[cdrkf]

https://forums.dolphin-emu.org/Thread-ryzen-1700x-stock-vs-dolphin-benchmark

Not to bad 10% lower then a 4770K meaning in this case Ryzen is below Haswell IPC but not by a massive margin

Better benchmark
http://www.legitreviews.com/wp-content/uploads/2017/03/dolphin-1700.jpg

I would ignore the first, because it is from one of those fake leaks released before launch. The performance reported in that leak wasn't confirmed by reviews.

And I would ignore the second because it is internally inconsistent. The 1700 and the 1800X are the same chip except for clocks, and the graph says that the 1700 OC to 4GHZ on all cores is slower that the 1800X at stock with 3.6GHz base and single core 4GHz turbo. A possible explanation of the discrepancy is that the 1800X is running with XFR activated in whose case the effective clocks aren't the clocks reported in the graph.

Here you have Dolphin benches from computerbase

Zen @3.6GHz performs like Haswell @3GHz in this bench.

Is that a single thread test per chance?

Edit: there is something strange about the scaling of that. If you look at the 1700- it's above a 3770k, which if we are talking base clocks has Ivy at a 500mhz advantage (3.5ghz base on the 3770 vs 3 on the 1700). If we are looking at turbo clocks then it's 3.9ghz vs 3.7ghz, however then your assertion that 'Zen @ 3.6ghz = Haswell @ 3ghz is wrong', because the 5960x turbos to 3.5ghz (so it would be 4 vs 3.5). The only thing that is strange about that though is the huge gap between Ivy and Haswell in this test, is this an AVX 2 enabled benchmark? Haswell didn't really have much ipc over Ivy in most test from what I remember...

It is not single thread. The huge gap between Haswell and Ivy in Dolphin is normal

@cdrkf

Its a synthetic benchmark thats why you cant really take it seriously. Probably most of what he has postet are synthetic ones.

It is not synthetic but a well-known emulator used by many people. I believe that jdwii uses it.

Juan, if that benchmark is Multi Threaded, then *how* is the 7700K the fastest chip, and how do parts in the same generation with less cores beat parts with much higher core counts? From what we can see, assuming the test is multi threaded in nature at all- it cannot use anything more than 8 threads. Which explains why there is no scaling on Intel going from 8 to 12 to 16 to 20 threads. It's all down to clock speed and little else.

A partially multi threaded test that uses an undetermined number of threads and scales on clock speed over core count isn't a very good way to test IPC on a 16 thread chip imo. We'd need to see where the 1500X sits in there (I'd also like to see some i5 and i3 parts in there to determine at what level core scaling becomes relevant).

 

[TMTOWTSAC]

@salgado18

I think by now everyone knows IPC isn't really a fixed quantity. It will change based on workloads and coding optimizations, hardware combinations, branch prediction, pipeline efficiency, cache hit misses, etc. But that doesn't seem to stop people from chasing down every obscure benchmark and use case scenario that favors one side or the other, then trying to apply it universally.

 

[jdwii]

@salgado18

I think by now everyone knows IPC isn't really a fixed quantity. It will change based on workloads and coding optimizations, hardware combinations, branch prediction, pipeline efficiency, cache hit misses, etc. But that doesn't seem to stop people from chasing down every obscure benchmark and use case scenario that favors one side or the other, then trying to apply it universally.

True IPC is different for every program but the issue is some reviews barely even test the processors correctly. Some don't even bother testing for IPC which is odd cause back in the old days they always compared a new design to several different designs. I have yet to see Ryzen compared with 15-20 benchmarks in IPC with several different architectures. I'm curious to how it compared to Bulldozer-Phenom-nehalem, sandy-kabylake.

IPC tests are pretty important since most people who need to upgrade are probably on 7 year old hardware.

 

[randomizer]

Mine is even older than that.

 

[jdwii]

Spoiler

https://forums.dolphin-emu.org/Thread-ryzen-1700x-stock-vs-dolphin-benchmark

Not to bad 10% lower then a 4770K meaning in this case Ryzen is below Haswell IPC but not by a massive margin

Better benchmark
http://www.legitreviews.com/wp-content/uploads/2017/03/dolphin-1700.jpg

I would ignore the first, because it is from one of those fake leaks released before launch. The performance reported in that leak wasn't confirmed by reviews.

And I would ignore the second because it is internally inconsistent. The 1700 and the 1800X are the same chip except for clocks, and the graph says that the 1700 OC to 4GHZ on all cores is slower that the 1800X at stock with 3.6GHz base and single core 4GHz turbo. A possible explanation of the discrepancy is that the 1800X is running with XFR activated in whose case the effective clocks aren't the clocks reported in the graph.

Here you have Dolphin benches from computerbase

Zen @3.6GHz performs like Haswell @3GHz in this bench.

Is that a single thread test per chance?

Edit: there is something strange about the scaling of that. If you look at the 1700- it's above a 3770k, which if we are talking base clocks has Ivy at a 500mhz advantage (3.5ghz base on the 3770 vs 3 on the 1700). If we are looking at turbo clocks then it's 3.9ghz vs 3.7ghz, however then your assertion that 'Zen @ 3.6ghz = Haswell @ 3ghz is wrong', because the 5960x turbos to 3.5ghz (so it would be 4 vs 3.5). The only thing that is strange about that though is the huge gap between Ivy and Haswell in this test, is this an AVX 2 enabled benchmark? Haswell didn't really have much ipc over Ivy in most test from what I remember...

It is not single thread. The huge gap between Haswell and Ivy in Dolphin is normal

@cdrkf

Its a synthetic benchmark thats why you cant really take it seriously. Probably most of what he has postet are synthetic ones.

It is not synthetic but a well-known emulator used by many people. I believe that jdwii uses it.

Juan, if that benchmark is Multi Threaded, then *how* is the 7700K the fastest chip, and how do parts in the same generation with less cores beat parts with much higher core counts? From what we can see, assuming the test is multi threaded in nature at all- it cannot use anything more than 8 threads. Which explains why there is no scaling on Intel going from 8 to 12 to 16 to 20 threads. It's all down to clock speed and little else.

A partially multi threaded test that uses an undetermined number of threads and scales on clock speed over core count isn't a very good way to test IPC on a 16 thread chip imo. We'd need to see where the 1500X sits in there (I'd also like to see some i5 and i3 parts in there to determine at what level core scaling becomes relevant).

He means it uses more then one core. Since intel isn't selling you a single core with their processor you have to use more then 4 cores before Ryzen will beat a Kaby-lake 4 core 8 threaded par(in gaming) processor and based on the lower clock speed and lower IPC you would probably have to have a game use 6 cores well before its even with a 7700K even more so with the infinity fabric hindering performance from ideal scaling like seen on a 6900K.

http://www.legitreviews.com/wp-content/uploads/2017/03/dolphin-1700.jpg

Despite all this i'll probably still jump back to Amd for awhile its not like the product won't work with my programs its just weaker. I basically just want to build a new AMD machine its been so long and my old Amd fanboy inside me is coming out more and more every day. I just wish we could all have a repeat of Amd K8

 

[salgado18]

Ok, I don't know what you think of Ars Technica and Overclockers.com, but they have good-looking graphs of single-core frequency-locked tests. The first thing I note is how close AMD has gotten to Intel, and the second is that lagging behind by so little is irrelevant right now. Let the other CPUs come out, Intel readjust their lineup (they have to), and the Zen core become normal in the marketplace, and then it makes sense to discuss small advantages from one or the other. But today, AMD has the upper hand in value.

(Edit: forgot to get the line that says all CPUs are locked at 3.5GHz here)
https://arstechnica.com/gadgets/2017/03/amd-ryzen-review/

http://www.overclockers.com/amd-ryzen-7-1800x-cpu-review/

 

[jdwii]

Its not so much in applications that's where Ryzen shines its more in games and in that they are basically behind haswell in IPC and by the looks of it by 10-15% in a lot of cases where the situation is CPU-bound.

Most game at 60hz and in that case even a I5 is a waste of money imo in most cases. 1400 quad core is plenty for that but for people who want to try and keep a 120+ frame rate its not in a lot of titles.

I5 has the same issue as Ryzen with high frame rate gaming that is why gamernexus and so forth say get a I7 for that task, but i can see potential with ryzen if the game is programmed for it. But with that its the same old wait and see approach from Amd that I've been hearing for about a decade.

R5 is a better deal then the I5 for gaming though and its actually a better gaming processor too.

 

[-Fran-]

How can you attribute the deficiencies in gaming only due to IPC and not just talk about "platform performance"?

It has been made *very* clear over the last couple of weeks that Ryzen has been getting (huge?) uplifts in *performance* around games when they actually account for Ryzen's new uArch via patches. Nothing to do with IPC alone.

PLEASE DO NOT KEEP USING "IPC" SO FREELY. What you're referring to, at least using the term as loosely as you are, is "performance". You can use IPC as *one* variant to attribute differences in *performance* across different configurations, but not *the sole problem behind* the differences.

Cheers!

 

[jdwii]

How can you attribute the deficiencies in gaming only due to IPC and not just talk about "platform performance"?

It has been made *very* clear over the last couple of weeks that Ryzen has been getting (huge?) uplifts in *performance* around games when they actually account for Ryzen's new uArch via patches. Nothing to do with IPC alone.

PLEASE DO NOT KEEP USING "IPC" SO FREELY. What you're referring to, at least using the term as loosely as you are, is "performance". You can use IPC as *one* variant to attribute differences in *performance* across different configurations, but not *the sole problem behind* the differences.

Cheers!

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

https://youtu.be/D_Yt4vSZKVk?t=271

Linus video on ram speed if anything Intel platforms seem to gain more then Amd ones for example in Rise of the Tomb raider

 

[-Fran-]

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

Ironically enough, no. IPC is actually the hardest metric to get from a program being pedantic. Do you actually know *how many instructions per cycle* Zen executes given a specific executable? Is that "140" number in CB just ADDs and MOVs sequences in a certain amount of seconds that we could dig up and give meaning to? I don't think anyone, not even the programmers behind those benchmarks, can accurately tell you how you can translate a score into proper "IPC" inside the CPU.

We just love to toss the term around when talking about just plain old "performance" when looking at benchmarks and making a huge assumption that IPC is the sole culprit for gains or losses when at the same time we know it is not the CPU alone. Ensue Jackie Chan meme at this point.

The numbers you get from benchmarks have IPC built into them, yes. Is it the most significant measurement? I have absolutely no idea. I would imagine for CB it is, since it might actually implement it's benchmarks around heavy usage of a few selected number of instructions and CPU paths, so the comparisons can be made valid there. Maybe a few other benchmarks keep that consistency across different revisions and versions. I'm pretty damn sure professional benchmarking tools do it when measuring specific things of a *platform*.

You have points of comparison for different numbers, but you can see external influence of the platform affecting those numbers, then *IPC* is not the correct term to use at all. Like palladin said a GOOD while ago, it's a deprecated term from RISC as a unit of measurement. Using it for CISC is weird at best if we are pedantic.

When people doesn't want to compromise on this, Hell breaks loose here. And yes, it is cyclic. Same when refusing to use the "-E" moniker to distinguish from Intel generational CPUs.

Cheers!

 

[salgado18]

Spoiler

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

Ironically enough, no. IPC is actually the hardest metric to get from a program being pedantic. Do you actually know *how many instructions per cycle* Zen executes given a specific executable? Is that "140" number in CB just ADDs and MOVs sequences in a certain amount of seconds that we could dig up and give meaning to? I don't think anyone, not even the programmers behind those benchmarks, can accurately tell you how you can translate a score into proper "IPC" inside the CPU.

We just love to toss the term around when talking about just plain old "performance" when looking at benchmarks and making a huge assumption that IPC is the sole culprit for gains or losses when at the same time we know it is not the CPU alone. Ensue Jackie Chan meme at this point.

The numbers you get from benchmarks have IPC built into them, yes. Is it the most significant measurement? I have absolutely no idea. I would imagine for CB it is, since it might actually implement it's benchmarks around heavy usage of a few selected number of instructions and CPU paths, so the comparisons can be made valid there. Maybe a few other benchmarks keep that consistency across different revisions and versions. I'm pretty damn sure professional benchmarking tools do it when measuring specific things of a *platform*.

You have points of comparison for different numbers, but you can see external influence of the platform affecting those numbers, then *IPC* is not the correct term to use at all. Like palladin said a GOOD while ago, it's a deprecated term from RISC as a unit of measurement. Using it for CISC is weird at best if we are pedantic.

When people doesn't want to compromise on this, Hell breaks loose here. And yes, it is cyclic. Same when refusing to use the "-E" moniker to distinguish from Intel generational CPUs.

Cheers!

IPC was an acceptable measurement up until Ryzen. We talked IPC of Bulldozer and derivatives, Sandy Bridge and derivatives, Jaguar, Conroe, K10 etc. The problem arises when we try to judge the entire performance of a CPU using just this one metric in a multi-threaded, SMT-enabled world.

All reviews show Ryzen going up and down on benchmarks, sometimes faster than Intel, sometimes slower, sometimes a match. Sometimes even the 7700k beats the 6850k. IPC is a good way of measuring the architecture in details, but every program will use resources in a different way, and that became very evident with Ryzen, because the performance is very close to Intel all around (instead of only sometimes, like Bulldozer).

So we could keep discussing IPC to pinpoint strenghts and weaknesses, like strong floating point but weak integer performance, but that can't be used to predict application performance. So, we can't say it's slower than this or that CPU without talking in which program.

All that being said, Ryzen is very competitive and viable against Intel right now. Let's dissect the architecture, but never forgetting how great it is.

 

[jdwii]

Spoiler

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

Ironically enough, no. IPC is actually the hardest metric to get from a program being pedantic. Do you actually know *how many instructions per cycle* Zen executes given a specific executable? Is that "140" number in CB just ADDs and MOVs sequences in a certain amount of seconds that we could dig up and give meaning to? I don't think anyone, not even the programmers behind those benchmarks, can accurately tell you how you can translate a score into proper "IPC" inside the CPU.

We just love to toss the term around when talking about just plain old "performance" when looking at benchmarks and making a huge assumption that IPC is the sole culprit for gains or losses when at the same time we know it is not the CPU alone. Ensue Jackie Chan meme at this point.

The numbers you get from benchmarks have IPC built into them, yes. Is it the most significant measurement? I have absolutely no idea. I would imagine for CB it is, since it might actually implement it's benchmarks around heavy usage of a few selected number of instructions and CPU paths, so the comparisons can be made valid there. Maybe a few other benchmarks keep that consistency across different revisions and versions. I'm pretty damn sure professional benchmarking tools do it when measuring specific things of a *platform*.

You have points of comparison for different numbers, but you can see external influence of the platform affecting those numbers, then *IPC* is not the correct term to use at all. Like palladin said a GOOD while ago, it's a deprecated term from RISC as a unit of measurement. Using it for CISC is weird at best if we are pedantic.

When people doesn't want to compromise on this, Hell breaks loose here. And yes, it is cyclic. Same when refusing to use the "-E" moniker to distinguish from Intel generational CPUs.

Cheers!

“How can you attribute the deficiencies in gaming only due to IPC and not just talk about "platform performance"?

Not 100% sure what you are talking about when it comes to platform performance you mean the whole platform that is a red herring all what matters is, are they using the same GPU are they using the same ram at the same frequency when testing software. I’d also argue that they should be using the top of the line platform like X370 but I doubt that makes any difference for CPU performance.

IPC+ Frequency is the only way to discuss single threaded performance

IPC by itself can mean different things but in most people’s, eyes its seen as performance per cycle. By that measure it will continue to be used freely in most people’s comments if you like we can say performance per cycle as that probably does make more sense.

 

[cdrkf]

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

Ironically enough, no. IPC is actually the hardest metric to get from a program being pedantic. Do you actually know *how many instructions per cycle* Zen executes given a specific executable? Is that "140" number in CB just ADDs and MOVs sequences in a certain amount of seconds that we could dig up and give meaning to? I don't think anyone, not even the programmers behind those benchmarks, can accurately tell you how you can translate a score into proper "IPC" inside the CPU.

We just love to toss the term around when talking about just plain old "performance" when looking at benchmarks and making a huge assumption that IPC is the sole culprit for gains or losses when at the same time we know it is not the CPU alone. Ensue Jackie Chan meme at this point.

The numbers you get from benchmarks have IPC built into them, yes. Is it the most significant measurement? I have absolutely no idea. I would imagine for CB it is, since it might actually implement it's benchmarks around heavy usage of a few selected number of instructions and CPU paths, so the comparisons can be made valid there. Maybe a few other benchmarks keep that consistency across different revisions and versions. I'm pretty damn sure professional benchmarking tools do it when measuring specific things of a *platform*.

You have points of comparison for different numbers, but you can see external influence of the platform affecting those numbers, then *IPC* is not the correct term to use at all. Like palladin said a GOOD while ago, it's a deprecated term from RISC as a unit of measurement. Using it for CISC is weird at best if we are pedantic.

When people doesn't want to compromise on this, Hell breaks loose here. And yes, it is cyclic. Same when refusing to use the "-E" moniker to distinguish from Intel generational CPUs.

Cheers!

Single thread test are a fair enough way to measure 'IPC' in relation to a specific workload (i.e. dividing performance by the clock speed to 'level the playing field' as it were). It isn't the whole story though.

I think it is worth noting though that Ryzen actually looks pretty strong from my perspective. I mean there's a lot of people complaining about 'poor' gaming performance, when in reality the only processor from Intel that can consistently best it is the 7700k, which is the undisputed best gaming cpu on the market at the moment. 8 core Zen hangs tight with the broadwell parts most of the time, there are a few cases where the performance is noticeably lower than you'd expect- although I'd argue those are edge cases (certainly the latest releases all look to perform well on Ryzen). It's certainly not in the position that FX was.

The only thing I find a little tiresome is a few people in this thread appear to be on a mission to discredit Ryzen by 'proving' the 'IPC is lower than Sandy Bridge' or other rather exaggerated claims. It's also worth noting that IPC alone isn't the only metric we should look at. Another important one is power consumption- irrespective of IPC, in workstation / production type workloads the Ryzen 8 core parts are *as fast in absolute performance* as an 8 core Broadwell E chip, whilst also offering as good / slightly better power consumption figures. The point is they have matched throughput *and* power consumption to Intels current high end platform. That is impressive, the fact it needs a few hundred extra mhz to achieve it is pretty much irrelevant if it can do so without the power draw going through the roof. That was really the biggest issue with FX- to get it to a point where it's performance was in line with Intel took huge amounts of power (if it was possible at all). Ryzen doesn't have that issue- the only 'problem' with the part is the process limits the upper clock speed to 4ghz, although I'd hardly call that a terrible issue either.

 

[jdwii]

Spoiler

IMO its the fairest way in AMD's favor to measure gaming performance as clock rate would be a losing battle with kaby-lake being able to OC to 4.7-5.0Ghz on so many machines.

IPC is the easiest measure when comparing the designs and was often the most common way. Gaming IPC on Ryzen is lower then programs that are more about throughoutput vs latency.

Ironically enough, no. IPC is actually the hardest metric to get from a program being pedantic. Do you actually know *how many instructions per cycle* Zen executes given a specific executable? Is that "140" number in CB just ADDs and MOVs sequences in a certain amount of seconds that we could dig up and give meaning to? I don't think anyone, not even the programmers behind those benchmarks, can accurately tell you how you can translate a score into proper "IPC" inside the CPU.

We just love to toss the term around when talking about just plain old "performance" when looking at benchmarks and making a huge assumption that IPC is the sole culprit for gains or losses when at the same time we know it is not the CPU alone. Ensue Jackie Chan meme at this point.

The numbers you get from benchmarks have IPC built into them, yes. Is it the most significant measurement? I have absolutely no idea. I would imagine for CB it is, since it might actually implement it's benchmarks around heavy usage of a few selected number of instructions and CPU paths, so the comparisons can be made valid there. Maybe a few other benchmarks keep that consistency across different revisions and versions. I'm pretty damn sure professional benchmarking tools do it when measuring specific things of a *platform*.

You have points of comparison for different numbers, but you can see external influence of the platform affecting those numbers, then *IPC* is not the correct term to use at all. Like palladin said a GOOD while ago, it's a deprecated term from RISC as a unit of measurement. Using it for CISC is weird at best if we are pedantic.

When people doesn't want to compromise on this, Hell breaks loose here. And yes, it is cyclic. Same when refusing to use the "-E" moniker to distinguish from Intel generational CPUs.

Cheers!

Single thread test are a fair enough way to measure 'IPC' in relation to a specific workload (i.e. dividing performance by the clock speed to 'level the playing field' as it were). It isn't the whole story though.

I think it is worth noting though that Ryzen actually looks pretty strong from my perspective. I mean there's a lot of people complaining about 'poor' gaming performance, when in reality the only processor from Intel that can consistently best it is the 7700k, which is the undisputed best gaming cpu on the market at the moment. 8 core Zen hangs tight with the broadwell parts most of the time, there are a few cases where the performance is noticeably lower than you'd expect- although I'd argue those are edge cases (certainly the latest releases all look to perform well on Ryzen). It's certainly not in the position that FX was.

The only thing I find a little tiresome is a few people in this thread appear to be on a mission to discredit Ryzen by 'proving' the 'IPC is lower than Sandy Bridge' or other rather exaggerated claims. It's also worth noting that IPC alone isn't the only metric we should look at. Another important one is power consumption- irrespective of IPC, in workstation / production type workloads the Ryzen 8 core parts are *as fast in absolute performance* as an 8 core Broadwell E chip, whilst also offering as good / slightly better power consumption figures. The point is they have matched throughput *and* power consumption to Intels current high end platform. That is impressive, the fact it needs a few hundred extra mhz to achieve it is pretty much irrelevant if it can do so without the power draw going through the roof. That was really the biggest issue with FX- to get it to a point where it's performance was in line with Intel took huge amounts of power (if it was possible at all). Ryzen doesn't have that issue- the only 'problem' with the part is the process limits the upper clock speed to 4ghz, although I'd hardly call that a terrible issue either.

I personally saw no one claim IPC is lower than sandy-bridge but lower then Haswell maybe you can give a quote of someone doing this? I have seen people blow off IPC or even tell people that 1700 is even worth mentioning if all you do is gaming and nothing else which is false. Plus all over the web I see people downplaying high frame-rate gaming something that is quite possibly wrong for the tech community to do.

You are right the 7700K from a pure gaming perspective is the best gaming processor on the market. Ryzen R5 series in most gaming tests beat a I5 even when the I5 is clocked at 4.7Ghz but overall its similar like how R7 is like the I5 in gaming tests. R3 I expect to be better then the I3 in gaming overall too.
Also I agree with the power consumption figures its quite impressive for Amd’s perspective Ryzen isn’t a terrible gaming processor that is not what I or Juan or anyone that I’ve seen so far on this thread has been saying its just not as good as the 7700K. To call Ryzen a bad gaming processor would mean one thought the I5 is a bad gaming processor which simply isn’t true.

 

[salgado18]

The point is they have matched throughput *and* power consumption to Intels current high end platform. That is impressive, the fact it needs a few hundred extra mhz to achieve it is pretty much irrelevant if it can do so without the power draw going through the roof. (...) the only 'problem' with the part is the process limits the upper clock speed to 4ghz, although I'd hardly call that a terrible issue either.

Most people forget (myself included) that the 6900k can be overclocked up to (or over) 4.0 GHz, pushing performance way higher than Ryzen... but also power consumption and heat. The fact that Ryzen is so efficient is a huge win for AMD.

But not going over 4 GHz is a drawback, not because of the Ryzen 7, but the 5's and 3's. They have lower core counts, they needed the push in clocks to be more competitive.