Interesting, I'm following the thread.

Quote from: william blake on February 10, 2020, 21:18:53

Quote from: Valantar on February 09, 2020, 18:25:22
lol. SPEC is an industry standard benchmark suite consisting of sample workloads from a wide range of real-world applications.

i know what spec is.

It sure didn't seem like it based on that first response.

Quote from: william blake on February 10, 2020, 21:18:53

Quote from: Valantar on February 09, 2020, 18:25:22
Talking about overall gaming performance across architectures on the other hand is of course possible

really? thank you. i am talking about overall performance(per mghz) and i am so tired arguing with people who think that ipc is a real number, somehow different from overall performance per mghz and somehow ipc is more important than overall performance per mghz.

I never said any of that. IPC is an attempt at describing the performance (instructions processed) per clock cycle (not MHz, as the term applies just as well to chips running at KHz or even Hz speeds) for a single component (such as a CPU), normally calculated from averaged results of a widely agreed upon set of benchmarks specifically stressing the part in question in various ways meant to represent a broad and representative selection of its capabilities. Some of these will always stress other parts of the system to various degrees - such as RAM or storage - but ideally only components that do not directly perform computational tasks but rather support the part in question. As GPUs do their own computations they don't fit this criteria, and as they need a driver to interface with the CPU properly you're adding several uncontrollable variables to the test. This effectively invalidates testing for IPC in GPU-reliant scenarios as you're running a mixed workload that stresses two parts and not just one, and with many interwoven processes dependent upon each other (cpu code, gpu code, gpu driver). This complexity means you can't simply run a series of simple tests and arrive at a comparable average, as the average then wouldn't tell you where the relevant bottlenecks are (in the CPU, GPU, or something in between?). There's also the issue of test normalization: should GPU clocks also be controlled, even if what one is looking for is CPU gaming performance? If so we would also need to agree upon a fixed reference GPU for testing, with a fixed set of specifications. And given that there are two major GPU vendors, this becomes rather problematic as testing done on one isn't necessarily transferable to the other. The more variables you add, the more you move away from relevant and realistic results, and you still aren't achieving the main goal of the term IPC, namely comparability. This is why using the term "gaming IPC" makes no sense - there's no reliable way to test such a thing, and even if there was the results wouldn't be comparable in the way single-component IPC comparisons are. And trying to make it so is just part of the massive abuse of the term IPC that we've seen over the past few years as use of the term has taken off. While there is indeed value to doing clock-normalized game testing on both CPUs and GPUs, none of this can easily be termed "IPC". "Performance" or "clock normalized performance" is much more suitable.

Quote from: Valantar on February 09, 2020, 18:25:22
lol. SPEC is an industry standard benchmark suite consisting of sample workloads from a wide range of real-world applications.

i know what spec is.

Quote from: Valantar on February 09, 2020, 18:25:22
Talking about overall gaming performance across architectures on the other hand is of course possible

really? thank you. i am talking about overall performance(per mghz) and i am so tired arguing with people who think that ipc is a real number, somehow different from overall performance per mghz and somehow ipc is more important than overall performance per mghz.

Quote from: william blake on February 01, 2020, 00:07:12
thanks a lot but i am not a spec user. my ipc consists of games, browsers and some video editing.

"SPEC user" - there's no such thing, lol. SPEC is an industry standard benchmark suite consisting of sample workloads from a wide range of real-world applications. Maybe look into what you're commenting on before dismissing it?

Quote from: S.Yu on February 01, 2020, 09:05:24

Quote from: Valantar on January 31, 2020, 23:30:48
Comet Lake is still Skylake, right? If so, Zen 2 is ahead on IPC, not slightly behind. Ahead by about 6.5% according to Anandtech's testing in SPEC2017.

IIRC they squeezed a few more percentages each year...or every couple year or so since Skylake, bottom line the IPC hasn't been at a total standstill since Skylake.

Actually it has. Beyond hardware mitigations replacing software fixes for security breaches, there have been zero relevant architectural changes from Skylake to Coffee Lake, and IPC is identical, with the only performance increases coming from clock speed increases. The testing I referred to used a 9900K, btw, so the 6% advantage is up-to-date.

Quote from: william blake on February 02, 2020, 02:22:34

Quote from: S.Yu on February 01, 2020, 21:12:55
GB is not specific enough about what it measures

is this https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf not enough?
and why it is worse than this https://www.spec.org/cpu2017/Docs/overview.html ?
anyway, iirc, spec(average from many tests) is pretty close to cinebench(single test) in terms of different architectures ipc comparison, but, ips measured in games is different to spec and cinebench.
and here is my main problem with using spec results as a reference of ipc. for a none-gamer its fine. but we should use something like (spec+games)/2, for and average pc user, it should provide a more accurate picture.

Talking about IPC in a gaming context is sadly almost impossible (or at least irrelevant), as adding a GPU inherently introduces too many uncontrollable variables to be able to identify something that can reliably be pointed out as CPU IPC. Different CPU architectures can treat various parts of the GPU driver differently, loading data differently, etc. - and this will likely even vary across GPU vendors too. Drivers will also have different levels of optimization for different architectures. So for any type of normalized test you'd need not only a repeatable workload (which can be done) and a selection of CPUs to test at a common clock speed, but also to test with a normalized GPU at a fixed performance level - but that becomes problematic as performance parity across CPU vendors with GPUs from different vendors can't be guaranteed. I.e. you'd end up with at least four classes: AMD GPU and AMD CPU, AMD GPU and Intel CPU, Nvidia GPU and AMD CPU, and Nvidia GPU and Intel CPU. Each would in all likelihood give different results. Simplifying this into a number you can call "IPC" becomes impossible, as external uncontrollable factors like GPU drivers and their optimizations for specific CPU architectures would inherently skew the numbers, invalidating the benchmark - you'd no longer be testing CPU IPC, but GPU driver optimization instead. This is demonstrated rather beautifully by Intel having lower CPU IPC, but still winning slightly in gaming performance thanks to a combination of higher clock speeds and better optimizations for their architecture (not to meniton, of course that most games have relatively few high-performance threads, which somewhat nullifies the advantage of having more fast cores vs. fewer slightly faster cores).

Talking about overall gaming performance across architectures on the other hand is of course possible, as it isn't dependent on normalizing anything beyond the workload, and simply asks which hardware configuration performs the best in closer to real-world scenarios. This is where Intel currently has the upper hand even in the desktop segment, though it remains to be seen if this also holds true to the more power limited mobile segment given Zen2's superior efficiency and clock scaling at low power.

Quote from: S.Yu on February 01, 2020, 21:12:55
GB is not specific enough about what it measures

is this https://www.geekbench.com/doc/geekbench5-cpu-workloads.pdf not enough?
and why it is worse than this https://www.spec.org/cpu2017/Docs/overview.html ?
anyway, iirc, spec(average from many tests) is pretty close to cinebench(single test) in terms of different architectures ipc comparison, but, ips measured in games is different to spec and cinebench.
and here is my main problem with using spec results as a reference of ipc. for a none-gamer its fine. but we should use something like (spec+games)/2, for and average pc user, it should provide a more accurate picture.

Quote from: ariliquin on February 01, 2020, 00:50:40
Whats with the title?  Worlds first 6 core 8 thread cpu? As if this is an achievement over the worlds first 8 core 16 thread laptop CPU from AMD? Seriously Intel has lost this race and will continue to do so until they fix their manufacturing. The gap is only getting bigger until then...

9th gen CFL-R H processors are the first 8C/16T mobile processors.
btw is the 4800H faster than 9980HK? Just curious.

Quote from: william blake on February 01, 2020, 15:05:00

Quote from: S.Yu on February 01, 2020, 09:05:24
It's just that SPEC is a much better written benchmark than GB.

i wonder what is that even mean?
imagine rendering is my job. so i want to know ipc of a different cpu architectures for..lets say, cinema 4d, which i work in.
is spec also "a much better written" than cinebench r20 in this case?

What is a good benchmark? A good benchmark accurately measures certain metrics with sufficient discrimination. GB is not specific enough about what it measures nor did the old version sufficiently discriminate modern hardware. There's little data on the new version.
Cinebench is better for your use case, which is a niche use case.

Quote from: Freddell on February 01, 2020, 09:11:31
The worlds first 6 core 8 thread CPU is the POWER8 processor in the Power S814 server released in 2014 and discontinued in 2019.
Intel is 6 years behind!
Author needs to check facts before spreading false news!

I'm the author of this piece. Respectfully, what you said is inaccurate. The Power8 architecture supports 8 threads per core. There was a 6-core variant of the Power8 S814, but that supported 48 threads at a time on 6 cores. I urge you not to call pieces out as being fake news when they're not.

Quote from: S.Yu on February 01, 2020, 09:05:24
It's just that SPEC is a much better written benchmark than GB.

i wonder what is that even mean?
imagine rendering is my job. so i want to know ipc of a different cpu architectures for..lets say, cinema 4d, which i work in.
is spec also "a much better written" than cinebench r20 in this case?

Quote from: ariliquin on February 01, 2020, 00:50:40
Whats with the title?  Worlds first 6 core 8 thread cpu? As if this is an achievement over the worlds first 8 core 16 thread laptop CPU from AMD? Seriously Intel has lost this race and will continue to do so until they fix their manufacturing. The gap is only getting bigger until then... 

The i9-9880H says hi.

The worlds first 6 core 8 thread CPU is the POWER8 processor in the Power S814 server released in 2014 and discontinued in 2019.
Intel is 6 years behind!
Author needs to check facts before spreading false news!

Quote from: Valantar on January 31, 2020, 23:30:48
Comet Lake is still Skylake, right? If so, Zen 2 is ahead on IPC, not slightly behind. Ahead by about 6.5% according to Anandtech's testing in SPEC2017.

IIRC they squeezed a few more percentages each year...or every couple year or so since Skylake, bottom line the IPC hasn't been at a total standstill since Skylake.

Quote from: ariliquin on February 01, 2020, 00:47:14
The main difference between Ryzen and Intel is not the core count, its 7nm vurses 14nm which will give AMD the advantage on thermals the main issues today will all thin and light perfromance laptops. Whats the piint of buying high end laptops that throttle to 800mhz consistently regardless of  their cores or 1 minutes scores?

Again, Intel's 14nm compares with TSMC's 10nm. I remember debunking this before, on this very site. Also AMD's disadvantage remains in idle power management, for some reason.

Quote from: ariliquin on February 01, 2020, 00:50:40
Whats with the title?  Worlds first 6 core 8 thread cpu? As if this is an achievement over the worlds first 8 core 16 thread laptop CPU from AMD? Seriously Intel has lost this race and will continue to do so until they fix their manufacturing. The gap is only getting bigger until then... 

I think it's meant to be sarcastic, "*as if* this is an achievement".

Quote from: Dan Ridenhour on February 01, 2020, 03:37:39
In the ole days (I haven't paid attention I a while) i5 and i7 chips rolled off the same production process and the 'difference' was how much of the chip passed performance testing.   So i5's were often i7s with failed cores, etc. disabled.     This sounds like a similar setup...  a 6 core 12 thread chip where at least 2 of the cores passed multi-threading QA... leave them turned on and give them a new chip to roll out the door... just by changing their Q/A process and what they leave 'on'.

You missed how in the "new days"(as in, last week) Nvidia's rumored to release a variant of RTX2060 on a heavily cut down RTX2080 die.

Quote from: william blake on February 01, 2020, 05:38:38

Quote from: The Hardcard on February 01, 2020, 03:15:19
There has never been a CPU that performed better on SPEC, but worse on other tasks.  If you want to know how a CPU will perform on games, browsers, video, photos, office tasks and so on, you can look at the SPEC results.

we can say this about other tests. geekbench for example. i mean, ipc is always an approximation, no matter what we take. and i never heard of the consensus about "ipc=spec ipc.

It's just that SPEC is a much better written benchmark than GB.

Quote from: The Hardcard on February 01, 2020, 03:15:19
There has never been a CPU that performed better on SPEC, but worse on other tasks.  If you want to know how a CPU will perform on games, browsers, video, photos, office tasks and so on, you can look at the SPEC results.

we can say this about other tests. geekbench for example. i mean, ipc is always an approximation, no matter what we take. and i never heard of the consensus about "ipc=spec ipc.

In the ole days (I haven't paid attention I a while) i5 and i7 chips rolled off the same production process and the 'difference' was how much of the chip passed performance testing.   So i5's were often i7s with failed cores, etc. disabled.     This sounds like a similar setup...  a 6 core 12 thread chip where at least 2 of the cores passed multi-threading QA... leave them turned on and give them a new chip to roll out the door... just by changing their Q/A process and what they leave 'on'.

Quote from: william blake on February 01, 2020, 00:07:12

QuoteWhat exactly will the i5-10500H be up against, though?

this questions always baffles me. there is no such thing as up against.

QuoteComet Lake is still Skylake, right? If so, Zen 2 is ahead on IPC, not slightly behind. Ahead by about 6.5% according to Anandtech's testing in SPEC2017.

thanks a lot but i am not a spec user. my ipc consists of games, browsers and some video editing.

There has never been a CPU that performed better on SPEC, but worse on other tasks.  If you want to know how a CPU will perform on games, browsers, video, photos, office tasks and so on, you can look at the SPEC results.