Jump to content

Question about Ryzen and Coffeelake IPC

geg43

Hi, I'm in the mood to buy myself a brand new system and I'm deciding between Intel and Amd. After reading the reviews, I'm left wondering why hasn't anyone done a core for core, clock for clock comparison of the two? I.E. 1600X @ 4ghz vs 8700k @ 4ghz. No turbo, same ram, etc etc. Is ryzen architecturally weaker than coffee lake or is it just the turbo that's responsible for the better scores?  

Link to comment
Share on other sites

Link to post
Share on other sites

because it's irrelevant

Primary System

  • CPU
    Ryzen R6 5700X
  • Motherboard
    MSI B350M mortar arctic
  • RAM
    32GB Corsair RGB 3600MT/s CAS18
  • GPU
    Zotac RTX 3070 OC
  • Case
    kind of a mess
  • Storage
    WD black NVMe SSD 500GB & 1TB samsung Sata ssd & x 1TB WD blue & x 3TB Seagate
  • PSU
    corsair RM750X white
  • Display(s)
    1440p 21:9 100Hz
Link to comment
Share on other sites

Link to post
Share on other sites

No it's not. Price may be a factor but the 8400 is beating the 1600 in reviews. Even though it has no HT and is 20-30$ cheaper. I want to know why.

Link to comment
Share on other sites

Link to post
Share on other sites

4 minutes ago, geg43 said:

Hi, I'm in the mood to buy myself a brand new system and I'm deciding between Intel and Amd. After reading the reviews, I'm left wondering why hasn't anyone done a core for core, clock for clock comparison of the two? I.E. 1600X @ 4ghz vs 8700k @ 4ghz. No turbo, same ram, etc etc. Is ryzen architecturally weaker than coffee lake or is it just the turbo that's responsible for the better scores?  

Different designs, different workloads. Compare what you want to use it for and go from there.

Link to comment
Share on other sites

Link to post
Share on other sites

That's the thing. I do productivity and game a bit. I just want to know what's the IPC of each when everything is even

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, geg43 said:

That's the thing. I do productivity and game a bit. I just want to know what's the IPC of each when everything is even

The same as KL vs. Ryzen, which was roughly a 10% lead.

PSU Tier List | CoC

Gaming Build | FreeNAS Server

Spoiler

i5-4690k || Seidon 240m || GTX780 ACX || MSI Z97s SLI Plus || 8GB 2400mhz || 250GB 840 Evo || 1TB WD Blue || H440 (Black/Blue) || Windows 10 Pro || Dell P2414H & BenQ XL2411Z || Ducky Shine Mini || Logitech G502 Proteus Core

Spoiler

FreeNAS 9.3 - Stable || Xeon E3 1230v2 || Supermicro X9SCM-F || 32GB Crucial ECC DDR3 || 3x4TB WD Red (JBOD) || SYBA SI-PEX40064 sata controller || Corsair CX500m || NZXT Source 210.

Link to comment
Share on other sites

Link to post
Share on other sites

Do you have any links? I want to see for myself.

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, geg43 said:

No it's not. Price may be a factor but the 8400 is beating the 1600 in reviews. Even though it has no HT and is 20-30$ cheaper. I want to know why.

core for core clock for clock is irrelevant because nobody will ever down clock their CPU or disable cores in real life. why would you kill performance you bought. so it's an irrelevant stupid comparisson.

Primary System

  • CPU
    Ryzen R6 5700X
  • Motherboard
    MSI B350M mortar arctic
  • RAM
    32GB Corsair RGB 3600MT/s CAS18
  • GPU
    Zotac RTX 3070 OC
  • Case
    kind of a mess
  • Storage
    WD black NVMe SSD 500GB & 1TB samsung Sata ssd & x 1TB WD blue & x 3TB Seagate
  • PSU
    corsair RM750X white
  • Display(s)
    1440p 21:9 100Hz
Link to comment
Share on other sites

Link to post
Share on other sites

No it's not. I want to know how powerful amd's architecture is compared to intel's.

 

Why are you being so difficult? 

Link to comment
Share on other sites

Link to post
Share on other sites

I don't have links at hand, but the gap in performance is mostly just clocks. Ryzen is comparable to Haswell/Broadwell, and IPC on Intel's side hasn't changed since Skylake, so the gap there is pretty small.

 

As for the 8400 beating the 1600X, Intel's stronger single-core and IMC will do that in most few-core applications. I'm sure an 8400 could beat my 1700X in a lot of games.

"Do as I say, not as I do."

-Because you actually care if it makes sense.

Link to comment
Share on other sites

Link to post
Share on other sites

I see. My local computer store has both the 1600x and 8600k in stock for 300$ each and I'm having a tough time deciding between the two. Maybe I should wait for zen2?

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, geg43 said:

I see. My local computer store has both the 1600x and 8600k in stock for 300$ each and I'm having a tough time deciding between the two. Maybe I should wait for zen2?

Waiting doesn't have s guaranteed timeline not any measurable benefit at this point. I have a 1700x and gaming is just fine. I got it for productivity over anything Intel had to offer at it's price point.

 

You want to see real life results, then get on Google, Bing, Yahoo, AskJeeves, Dogpile, DuckDuckGo, or even YouTube do a quick search. I'm fairly certain you can find something.

 

Did you look before asking?

Link to comment
Share on other sites

Link to post
Share on other sites

Yes I did look, but there are no direct IPC comparisons. At best is estimations.

 

Man why is this community so belligerent and hostile? A man cant come and ask a simple question without people jumping down his throat?

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, geg43 said:

Yes I did look, but there are no direct IPC comparisons. At best is estimations.

I don't think there's a good way to not estimate it.

Here's a comparison with Kaby Lake, as said before it's fundamentally the same architecture as Coffee. It's from the somewhat early days, though, so Ryzen's performance has improved through the various updates and tweaks AMD put out.

I'm having trouble finding a current comparison, since everyone's already reviewed Ryzen and not very many people are still asking how the IPC compares now that it's been out for a while.

 

As a side note, you should quote people when you reply (arrow under the post), that way they get a notification.

"Do as I say, not as I do."

-Because you actually care if it makes sense.

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, geg43 said:

Hi, I'm in the mood to buy myself a brand new system and I'm deciding between Intel and Amd. After reading the reviews, I'm left wondering why hasn't anyone done a core for core, clock for clock comparison of the two? I.E. 1600X @ 4ghz vs 8700k @ 4ghz. No turbo, same ram, etc etc. Is ryzen architecturally weaker than coffee lake or is it just the turbo that's responsible for the better scores?  

Let's break this down:

 

1) CPU instructions are different in AMD and INTEL even those with identical names (AVX, AES, etc). They are different implementations of the same concept, not directly comparable, just like the "thrust" in electric and fuel engine cars.

2) How can you compare IPC between different CPU's with different instruction-set implementations?. Let's say AMD Ryzen has 10 AVX per cycle (no it does not, it's a simple way to put this). While Intel Covfefe lake has 11AVX instructions per cicle, does it mean intel is better? No. Different instructions set. There is no guide line to determine what should an AVX instruction do, nor how long should it take. Plus the impact in every architecture is way different (different amounts of cache, different cache speeds). So IPC is not a way to compare CPUs. it's obscure and most data is lost to overgeneralizing situations that are unknown to us (implementation secrets) and inherent complexity of the matter. 

Any real world application has mixed instruction sets involved, every instruction needs different "sizes" in the execute module. So you end up with a bunch of statistical data, that compares things that are not comparable in the first place. Worthless data.

3) Ryzen has architectural flaws. They are evident when you go up to threadripper (latency between distant cores in packs communicating through infinity fabric). 

Intel has them too, Haswel had an AVX bug that made it eat up much more current than it should. Plus it had TSX instruction set disabled due to random bugs. Skylake and Kaby lake have an Hyperthreading bug that is evident under certain loads and particularly problematic for servers... etc. There is no bug free hardware nor software.

Ryzen is stronger than intel in terms of scalability. Its a super scalar architecture similar to what NVidia did with Pascal. It lets them add more cores without making a lot of changes. Of course it has it's own limits, tightly linked to "Infinity fabric way of things". Intel has a simpler core design that allows them to archive faster clocks. All in all, the end user prefers faster clockspeed whenever possible, due to lack of software support for multicore monsters. Some games are made to run in 32 core monsters, but the vast majority is made for a few cores. (It makes programming even more difficult when you try to go past a handful of threads). So "weak" and "strong" are absurdly dumbed down terms to speak of architectural perks and flaws. 

4) Scores. Use them as guidelines not as the absolute truth. Today for pure gaming, having an intel quad-core or Hexa-core overclocked up to 5.0Ghz is the indisputable best way to go. It has nothing to do with hardware. It's pure software guilt here. Most games have design patters that link heavily some game related features (shadows, physics, input, etc) to one thread. Even if the O.S tries to spread the load, it simply can't. Why is it better you may ask?. Well, games need at least 2 years to be developed. Ryzen came out this year, so there is no game today that was built with ryzen in mind. Plus Zen hits a huge brickwall after 4.0Ghz due to core stability, Intel hits the same brickwall after 5.0Ghz due to heat. 

 

The key element to understand this Versus is understanding workload. If you deal with games, the "no compromises" CPU is intel. If you deal with budget, your best bet is Zen. You won't be winning any high score contest though. But this is a constant since the beginning. 

 

Link to comment
Share on other sites

Link to post
Share on other sites

14 hours ago, geg43 said:

No it's not. Price may be a factor but the 8400 is beating the 1600 in reviews. Even though it has no HT and is 20-30$ cheaper. I want to know why.

Why do you care? Also why do you need to see those reviews to know that? Intel has the same clockspeeds, but fewer threads and is matching it. Intel has better single core performance. Intel has better ipc. There. Logic did the work eather than reviewers. 


Main System: EVGA GTX 1080 SC, i7 8700, 16GB DDR4 Corsair LPX 3000mhz CL15, Asus Z370 Prime A, Noctua NH D15, EVGA GQ 650W, Fractal Design Define R5, 2TB Seagate Barracuda, 500gb Samsung 850 Evo
Secondary System: EVGA GTX 780ti SC, i5 3570k @ 4.5ghz, 16gb DDR3 1600mhz, MSI Z77 G43, Noctua NH D15, EVGA GQ 650W, Fractal Design Define R4, 3TB WD Caviar Blue, 250gb Samsung 850 Evo
 
Link to comment
Share on other sites

Link to post
Share on other sites

Why are u comparing two unevenly priced CPU's. It's not the architecture that matters, its price to performance. AMD is slightly worse, but give you more of it so it's not a fair comparison

CPU: Ryzen 5 1600 MOBO: MSI Tomahawk B350 GPU: Reference cooled GTX 980 Storage: Intel SSD5 256Gb RAM: 8gb Geil EVO Potenza Case:  Phanteks p300 PSU: EVGA 500 watt CPU Cooler: AMD wraith spire

 

 

Steam: maxarooni4

Battle.net: MAX

 

If you have an Oculus HMU in dead and buried   

Link to comment
Share on other sites

Link to post
Share on other sites

15 hours ago, SquintyG33Rs said:

because it's irrelevant

 

14 hours ago, SquintyG33Rs said:

core for core clock for clock is irrelevant because nobody will ever down clock their CPU or disable cores in real life. why would you kill performance you bought. so it's an irrelevant stupid comparisson.

 

It is nice to know architecturally how close AMD is, and this also give a rough estimate to the Ghz difference. as in a AMD at 4.0 is the same as Inlet at 3.6 which would help with understanding CPU's with out the need of finding benchmarks for that specific one.

if you want to annoy me, then join my teamspeak server ts.benja.cc

Link to comment
Share on other sites

Link to post
Share on other sites

15 hours ago, geg43 said:

I see. My local computer store has both the 1600x and 8600k in stock for 300$ each and I'm having a tough time deciding between the two. Maybe I should wait for zen2?

no, just get the 8600K then, it's quite a bit faster than the Ryzen CPU.

| CPU: Core i7-8700K @ 4.89ghz - 1.21v  Motherboard: Asus ROG STRIX Z370-E GAMING  CPU Cooler: Corsair H100i V2 |
| GPU: MSI RTX 3080Ti Ventus 3X OC  RAM: 32GB T-Force Delta RGB 3066mhz |
| Displays: Acer Predator XB270HU 1440p Gsync 144hz IPS Gaming monitor | Oculus Quest 2 VR

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, AMDPRO said:

Why are u comparing two unevenly priced CPU's. It's not the architecture that matters, its price to performance. AMD is slightly worse, but give you more of it so it's not a fair comparison

I know amd is worse than intel, but I want to know how close enough amd is to intel.

 

3 hours ago, Zeitec said:

Why do you care? Also why do you need to see those reviews to know that? Intel has the same clockspeeds, but fewer threads and is matching it. Intel has better single core performance. Intel has better ipc. There. Logic did the work eather than reviewers. 

Again core for core, clock for clock. I want to know the performance delta. When the first core 2 duo's were released that's all that tech sites could talk about IPC and performance per clock.

Link to comment
Share on other sites

Link to post
Share on other sites

19 hours ago, SquintyG33Rs said:

because it's irrelevant

It's sad to see someone say this as it compares the CPU architecture and can show how close Amd really is clock for clock which is 100% relevant used to be done all the time back in the day.

 

Basically coffee-lake has the same IPC as kaby-skylake for the past 3 generations Intel hasn't moved IPC. 

 

That said Ryzen is a good 10-15% behind on average and just about within Haswell IPC i tested this myself and compared my chip to my 4790K. Things that can uses AVX2 intel will see greater gains then 10-15% compared to Ryzen. 

 

  

Link to comment
Share on other sites

Link to post
Share on other sites

18 hours ago, SquintyG33Rs said:

core for core clock for clock is irrelevant because nobody will ever down clock their CPU or disable cores in real life. why would you kill performance you bought. so it's an irrelevant stupid comparisson.

No its not as some programs can only use 2-4 cores and it compares the CPU architecture for example how does the 1600X compare to the 8700K same core count at the same frequency? 

 

How far behind is Amd on a per core basis? Ryzen 3 series takes a 3.8Ghz OC to start to overcome a locked 4460. These things are interesting from a academic standpoint 

 

Edit if i had to say i would claim this 

 

Ryzen IPC is between ivy-haswell for gaming and latency sensitive tasks such as games and emulation probably due to CCX latency which the architecture will always have even with Ryzen III

Ryzen IPC is slightly above haswell for throughput based applications such as handbrake, CineBench and so on. 

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, jdwii said:

It's sad to see someone say this as it compares the CPU architecture and can show how close Amd really is clock for clock which is 100% relevant used to be done all the time back in the day.

 

Basically coffee-lake has the same IPC as kaby-skylake for the past 3 generations Intel hasn't moved IPC. 

 

That said Ryzen is a good 10-15% behind on average and just about within Haswell IPC i tested this myself and compared my chip to my 4790K. Things that can uses AVX2 intel will see greater gains then 10-15% compared to Ryzen. 

 

  

I explained my comment further down.... Sure it's neat to know but it doesn't hold value in real usage. Therefore it is irrelevant. Also people have to stop recomenting on the first reply to a thread when there is litterally already over 100 replies over the past day

Primary System

  • CPU
    Ryzen R6 5700X
  • Motherboard
    MSI B350M mortar arctic
  • RAM
    32GB Corsair RGB 3600MT/s CAS18
  • GPU
    Zotac RTX 3070 OC
  • Case
    kind of a mess
  • Storage
    WD black NVMe SSD 500GB & 1TB samsung Sata ssd & x 1TB WD blue & x 3TB Seagate
  • PSU
    corsair RM750X white
  • Display(s)
    1440p 21:9 100Hz
Link to comment
Share on other sites

Link to post
Share on other sites

45 minutes ago, geg43 said:

I know amd is worse than intel, but I want to know how close enough amd is to intel.

 

Again core for core, clock for clock. I want to know the performance delta. When the first core 2 duo's were released that's all that tech sites could talk about IPC and performance per clock.

Okay if you really want a useless arbitrary statistic, the 8400 goes up to 4vhz turbo. The 1800x does at well. Look at single core performance between those 2.


Main System: EVGA GTX 1080 SC, i7 8700, 16GB DDR4 Corsair LPX 3000mhz CL15, Asus Z370 Prime A, Noctua NH D15, EVGA GQ 650W, Fractal Design Define R5, 2TB Seagate Barracuda, 500gb Samsung 850 Evo
Secondary System: EVGA GTX 780ti SC, i5 3570k @ 4.5ghz, 16gb DDR3 1600mhz, MSI Z77 G43, Noctua NH D15, EVGA GQ 650W, Fractal Design Define R4, 3TB WD Caviar Blue, 250gb Samsung 850 Evo
 
Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, Zeitec said:

Okay if you really want a useless arbitrary statistic, the 8400 goes up to 4vhz turbo. The 1800x does at well. Look at single core performance between those 2.

"Arbitrary"  

based on random choice or personal whim, rather than any reason or system.

 

It's not useless as one see how far apart they are and how much Amd needs to improve their IPC to become even with Intel right now its around 10-15%. 

 

If ryzen was a good 30% behind intel in IPC i wouldn't have bought into the platform as i could never see them catching up instead they are pretty close meaning future Ryzen revisions can make Amd basically even with Intel in terms of IPC and future fabrication can allow for higher frequency's

 

The point that you claim its irreverent is the main reason why tech reviewers need to include it in their tests.  

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×