-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Ryzen "2" ?
- Thread starter beany_bot
- Start date
Caporegime
No MT.
Permabanned
Caporegime
Like I've said in the past, you can't really apply IPC to that comparison. In the same way people don't say an i7 has higher ipc of an i5 in the same generation etc.
The performance is obviously the most important factor though.
Caporegime
bah we need him to run us ST to compare
Its a few %
Why not? your reasoning in that is flawed, the i7 has 8 threads the i5 only 4, the 1600 and 8700K have same number of threads, 12.
Caporegime
An FX8350 has 8 threads, an i7 2600K has 8 threads, in Cinebench R15 in multithreaded they'd probably be nigh on equal, yet no one would ever say they had similar IPC.
Caporegime
The 2600K and FX8350 are not "multithreaded they'd probably be nigh on equal" the 2600K is clearly faster, and the FX8350 is 8 cores 8 threads vs 4 cores 8 thread on the 2600K... completely different.
Permabanned
Does someone not understand what IPC is or something ??
Easy version is how Much POWER/MHZ you get
You can have 5ghz cpu lets say Sandy get beat by 4.8 Ivy because Isy got more IPC.
Ryzens IPC is very close to Intels. Thing is Ryzen does not overclock to 5ghz thats why its slower. If rysen would OC to 5ghz it would CRUSH COMPETITION like Raja said
gotta love this lie
Easy version is how Much POWER/MHZ you get
You can have 5ghz cpu lets say Sandy get beat by 4.8 Ivy because Isy got more IPC.
Ryzens IPC is very close to Intels. Thing is Ryzen does not overclock to 5ghz thats why its slower. If rysen would OC to 5ghz it would CRUSH COMPETITION like Raja said
gotta love this lie
Caporegime
You're overestimating the 2600K. In 8 threaded apps its still a 4 core CPU, the FX8350's extra cores would start to either be close or edge it out.
I have no problems with stating that at the same clocks in multi threaded situations a 1600 can match an 8700K. I have a problem with the use of IPC, as it doesn't really marry up given that unless you're using those SMT/HT threads the SMT superiority doesn't matter because it's not getting used, so you're going to be at the mercy of the speed of the cores and their IPC.
Like I say, the overall performance is the most important factor. But so's reasoning behind it.
Last edited:
Permabanned
But FX8350 was NOT A TRUE 8 core
https://www.kitguru.net/components/...d-responds-to-bulldozer-class-action-lawsuit/
It's not 8 core its 2 modules with 4 cores
Yes I REMEMBER EVERYTHING
corrected after memory refresh hehe
Last edited:
Caporegime
They are a completely different layout Martini, one is a 4 core 8 thread the other an 8 core 8 thread.
The 1600 and 8700K are both 6 core 12 thread.
The 1600 and 8700K are both 6 core 12 thread.
Permabanned
Caporegime
If we're using the phrase IPC to compare 6C/12T to 6C/12T then surely any i7 core's IPC is compared at 1C/2T to any core regardless if it can only execute one thread or not.
Which is why I think it's a fallacy.
Agree to disagree.
Permabanned
Calculation of IPC
The number of instructions per second and floating point operations per second for a processor can be derived by multiplying the number of instructions per cycle with the clock rate (cycles per second given in Hertz) of the processor in question. The number of instructions per second is an approximate indicator of the likely performance of the processor.
The number of instructions executed per clock is not a constant for a given processor; it depends on how the particular software being run interacts with the processor, and indeed the entire machine, particularly the memory hierarchy. However, certain processor features tend to lead to designs that have higher-than-average IPC values; the presence of multiple arithmetic logic units (an ALU is a processor subsystem that can perform elementary arithmetic and logical operations), and short pipelines. When comparing different instruction sets, a simpler instruction set may lead to a higher IPC figure than an implementation of a more complex instruction set using the same chip technology; however, the more complex instruction set may be able to achieve more useful work with fewer instructions.
The number of instructions per second and floating point operations per second for a processor can be derived by multiplying the number of instructions per cycle with the clock rate (cycles per second given in Hertz) of the processor in question. The number of instructions per second is an approximate indicator of the likely performance of the processor.
The number of instructions executed per clock is not a constant for a given processor; it depends on how the particular software being run interacts with the processor, and indeed the entire machine, particularly the memory hierarchy. However, certain processor features tend to lead to designs that have higher-than-average IPC values; the presence of multiple arithmetic logic units (an ALU is a processor subsystem that can perform elementary arithmetic and logical operations), and short pipelines. When comparing different instruction sets, a simpler instruction set may lead to a higher IPC figure than an implementation of a more complex instruction set using the same chip technology; however, the more complex instruction set may be able to achieve more useful work with fewer instructions.
Permabanned
Anyhow I think Zen+ looks Good for what it IS an Evolution of Zen. I think AMD is making mistake calling it Ryzen 2 should stick to Ryzen Plus or Ryzen Refresh. So people would not get tooo hyped.
but its AMD masters of HYPEEEE TRAINNN that usually derails
but its AMD masters of HYPEEEE TRAINNN that usually derails
This really makes no sense. How is comparing "SMT A" to "the lack of SMT A" in any way similar to comparing "SMT A" to "SMT B"?
Maybe the term "IPC" wasn't used but everyone made comparisons between Core i5s and Core i7s, with SMT obviously being the differentiator. IPC doesn't inherently mean "per core", although that's also a legitimate comparison. IPC is just that, instructions per clock, and you can do more if the application uses SMT effectively.
When most people use the term IPC they do really mean "IPC per core" though. It's an odd one because a "core" constitutes lots of parts, and you wouldn't say "it's not a valid IPC comparison because one chip has more cache" or "it's not a valid IPC comparison because one chip has AVX512 instruction set and the other doesn't". So I feel SMT is a perfectly valid variable to include, it just obviously makes no difference to single threaded workloads.
Last edited:
Caporegime
Maybe I don't like change.
I didn't try to deny people comparing none HT chips against HT chips, but it wasn't ever called an IPC difference.
I guess it's because the relative IPC difference between Ryzen and Coffeelake will change with the amount of cores/threads being used in an application.
There could be a 10% difference between the two at the same clocks in an application (Using 6 cores, but not threaded enough to use the SMT/HT threads). That application could get an update to utilise more threads and then it could drop further (Bigger difference depending on amount of threads as SMT nets higher gain per thread). Nothing physical has changed but the perceived IPC in such a comparison would have.
Last edited:
Caporegime
Sorry to be pedantic, but it's not limited to just those singlethread. It could use 6 cores and it would be classified as rather multithreaded but the SMT superiority would be nowhere to be found.
Associate
IPC has always been a vague perceptual hand-wave and it always varied massively for many reasons, primarily instruction set and cache sizes. SMT vs HT is the least of your worries if you take IPC as a serious measure of anything.