lets start digging the i7-8700K grave

[TheV_Machine]

i Hope im not getting kicked into the ass linking another tech-yt´s vid on ltt forums, but jump to 51:00 ( 51:00 in fahrenheit inches ) for the news on the leaked i7.8700K benchmarks. If these numbers are correct, so that this benchmarked cpu really was a full funtional ready-to-sell 8700K, then he can already start putting himself into the grave. Im adding a screenshot of my ryzen 5 1600 cpu cinebench scores. Bare in mind, my ryzen currently has a 3.6 ghz oc(planning to go higher) and 3200 mhz 8gb dual channel ram kit on a ASUS Prime B350M-A mobo. my ryzen 5 1600 (not oc´ed to its full capability) scored 1209, while the 8700K at stock scored 1230. If the mentioned price increase in comparison to the ryzen 5 1600 really is about 100 bucks, as kyle and paul mention it in the video, thats really really reeeeeeally bad. if this is what intels plans to counter amd with, i am now very happy that i didnt wait for coffee lake to drop...

[MoonlightSylv]

It's most likely still going to be quite a bit faster in single threaded applications. So not dead yet.

[ARikozuM]

People are putting too much stock in a synthetic benchmark. 

[tp95112]

I dont really care until real benchmarks come out.

[TheV_Machine]

4 minutes ago, ShadowTechXTS said:

It's most likely still going to be quite a bit faster in single threaded applications. So not dead yet.

im not really comfortable paying up to 120 bucks more for a better single core performance when most programs in our digitalised world start utilizing more cores

 

[hconverse02]

Well if you watch a) linus's ryzen right for you video and any other youtuber's comparisons of ryzen and intel, you will find that this is the case with multithreaded performances.  They perform very similarly but intel always has dominated in single threaded and (most) gaming performances.

[TheV_Machine]

1 minute ago, ARikozuM said:

People are putting too much stock in a synthetic benchmark. 

agreed, but you have to admit they give atleast a indicator where the performance of a product is going to be

[MoonlightSylv]

1 minute ago, TheV_Machine said:

im not really comfortable paying up to 120 bucks more for a better single core performance when most programs in our digitalised world start utilizing more cores

Valid point. Although it depends on the customer. I'm content with my 1600.

[TheV_Machine]

3 minutes ago, hconverse02 said:

Well if you watch a) linus's ryzen right for you video and any other youtuber's comparisons of ryzen and intel, you will find that this is the case with multithreaded performances.  They perform very similarly but intel always has dominated in single threaded and (most) gaming performances.

only becaus most games still use around 4-6 cores at most to run, but thats starting to change and these cap of 4-6 cores was the benfit for smoother gaming experience on the intel platform caused by the better single core performance. the advantages of having better single core performance are groing smaller but still slowly. coffee lake might not be affected by that, but the generation after that should

[ARikozuM]

7 minutes ago, TheV_Machine said:

agreed, but you have to admit they give atleast a indicator where the performance of a product is going to be

It doesn't. Multi-threaded performance isn't the same as single-threaded leveraging multiple cores. One task can be split (multi-thread) and the other cannot (single-thread). BF1 is a prime example of single-threaded performance leveraging multiple cores. 

[TheRandomness]

-Thread moved to CPUs, Motherboards, and Memory. And since the topic sounds like a "let's hate Intel" topic, it's one post away from getting locked.

[TheV_Machine]

3 minutes ago, ARikozuM said:

It doesn't. Multi-threaded performance isn't the same as single-threaded leveraging multiple cores. One task can be split (multi-thread) and the other cannot (single-thread). BF1 is a prime example of single-threaded performance leveraging multiple cores. 

cant tell from my own experience, i dont have bf1 and never played it anywhere else, but this single threaded leveraging multiple cores wont save intels face for a whole new generation considering the mentioned pricetag, if this price tag becomes reality, simply because as i said earlier, real mulit threaded workload is advancing and i know it seems that im using this as an universal excuse to talk shit about the 8700K, but what will single core performance be useful at if multi threaded workload becomes universal, maybe not when coffee lake hits the stores, but after that it gonna be interesting to see

[Drak3]

26 minutes ago, TheV_Machine said:

im not really comfortable paying up to 120 bucks more for a better single core performance when most programs in our digitalised world start utilizing more cores

 

Multicore performance = (Single core performance x Core count) - Scaling

 

Given a 5% IPC advantage for Sky and up. Given a worst case scenario of 4.8GHz average OC (6700K avg), and best case of 5GHz (7700K avg) for the 8700K, and 3.9GHz average for non-X Ryzen SKUs, we can easily get an idea of relative performance.

 

 

Ryzen will be considered 100% for this math. Turbo clocks not considered. Performance will be reported simply as the GHz value + IPC advantage (in case of 8700K) 

 

Ryzen 5 1600 at stock (3.2GHz), 6 cores. Theoretical performance: 19.2GHz

Ryzen 5 1600 at avg. OC (3.9GHz), 6 cores. Theoretical performance: 23.4GHz

Ryzen 7 1800X at avg. OC (4GHz), 8 cores. Theoretical performance: 32GHz.

 

8700K at 6700K avg OC (4.8GHz), 6 cores, +5% IPC advantage. Theoretical performance: 30.24 GHz

8700K at 7700K avg OC (5GHz), 6 cores, +5% IPC advantage. Theoretical performance: 31.5GHZ

 

 

But that is also assuming the disregard of AVX, where the 1800X @ 4GHz is in the same ballpark as the 7700K @ 4.2GHz. That is ignoring multicore scaling (likely to get worse as you add cores). That is ignoring multitasking with multiple single thread applications.

 

The 8700K, worst case, is still well above a Ryzen 6 core. Best case, it's right behind Ryzen's best offering.

[ARikozuM]

Just now, TheV_Machine said:

cant tell from my own experience, i dont have bf1 and never played it anywhere else, but this single threaded leveraging multiple cores wont save intels face for a whole new generation considering the mentioned pricetag, if this price tag becomes reality, simply because as i said earlier, real mulit threaded workload is advancing and i know it seems that im using this as an universal excuse to talk shit about the 8700K, but what will single core performance be useful at if multi threaded workload becomes universal, maybe not when coffee lake hits the stores, but after that it gonna be interesting to see

You'll never be able to split certain functions without adding latency (pauses) or time-out errors. Gaming isn't as easy as Premiere or Autodesk when it comes to multi-threading. There are too many functions that have to be done quickly and in-sequence. AI functionality can be split, but that causes issues in itself if there has to be a check for changes every other second based on what the player does between those checks. 

 

There will always be the consumer line-up for those who only do web browsing and gaming. If you're a content creator or do 3D modeling (CAD), you're more likely to build a workstation system with high core counts or buy from a reputable vendor that put the time to verify components (Puget) rather than DIY. 

[vanished]

Something about the 8700k?  OK  

tbh, I'm excited.  So far it's expected to be similar speed to a 7700k, similar price (~$350), and yet bumped up from a 4 to a 6 core.  If that's true, not only will it continue to dominate in single-threaded situations, but it will even beat the 1800X in multi-threaded.

[TahoeDust]

lol @ thinking R5 is going to compete with the 8700k because a overclocked R5 scores almost as high in one synthetic benchmark.

[TheV_Machine]

14 minutes ago, Drak3 said:

Multicore performance = (Single core performance x Core count) - Scaling

 

Given a 5% IPC advantage for Sky and up. Given a worst case scenario of 4.8GHz average OC (6700K avg), and best case of 5GHz (7700K avg) for the 8700K, and 3.9GHz average for non-X Ryzen SKUs, we can easily get an idea of relative performance.

 

 

Ryzen will be considered 100% for this math. Turbo clocks not considered. Performance will be reported simply as the GHz value + IPC advantage (in case of 8700K) 

 

Ryzen 5 1600 at stock (3.2GHz), 6 cores. Theoretical performance: 19.2GHz

Ryzen 5 1600 at avg. OC (3.9GHz), 6 cores. Theoretical performance: 23.4GHz

Ryzen 7 1800X at avg. OC (4GHz), 8 cores. Theoretical performance: 32GHz.

 

8700K at 6700K avg OC (4.8GHz), 6 cores, +5% IPC advantage. Theoretical performance: 30.24 GHz

8700K at 7700K avg OC (5GHz), 6 cores, +5% IPC advantage. Theoretical performance: 31.5GHZ

 

 

But that is also assuming the disregard of AVX, where the 1800X @ 4GHz is in the same ballpark as the 7700K @ 4.2GHz. That is ignoring multicore scaling (likely to get worse as you add cores). That is ignoring multitasking with multiple single thread applications.

 

The 8700K, worst case, is still well above a Ryzen 6 core. Best case, it's right behind Ryzen's best offering.

Seriously, im appreciating the effort you Put into Your comment, but considering scaling as a variable in real world Performance, this variable ist to important to call the conclusion you draw well-based, or at least thats my opinion

 

15 minutes ago, ARikozuM said:

You'll never be able to split certain functions without adding latency (pauses) or time-out errors. Gaming isn't as easy as Premiere or Autodesk when it comes to multi-threading. There are too many functions that have to be done quickly and in-sequence. AI functionality can be split, but that causes issues in itself if there has to be a check for changes every other second based on what the player does between those checks. 

 

There will always be the consumer line-up for those who only do web browsing and gaming. If you're a content creator or do 3D modeling (CAD), you're more likely to build a workstation system with high core counts or buy from a reputable vendor that put the time to verify components (Puget) rather than DIY. 

"Never" isnt a word i use often in consumer electronics considering the Last decade of development but you Just thought me something about gaming applications and why there are dufficulties to improve their multi thread Performance. Still i will stay with my Point that the cinebench score give atleast an indication of multi threaded workload and that improvent between the Stock 8700K and the 3.6 GHz oc'ed 1600 doesnt fit with a 100 bucks more in my opinion

[TahoeDust]

24 minutes ago, Drak3 said:

The 8700K, worst case, is still well above a Ryzen 6 core. Best case, it's right behind Ryzen's best offering.

...and that is only considering multicore performance.  When you start factoring in the huge single core advantage the 8700k has, suddenly Ryzen is not looking so great.

[TheV_Machine]

18 minutes ago, Ryan_Vickers said:

Something about the 8700k?  OK  

tbh, I'm excited.  So far it's expected to be similar speed to a 7700k, similar price (~$350), and yet bumped up from a 4 to a 6 core.  If that's true, not only will it continue to dominate in single-threaded situations, but it will even beat the 1800X in multi-threaded.

Based on These leaked scores, im far from seeing the 1800X beaten in multi threaded workload, but i might Just underestimate the oc scaling of the 8700K

[vanished]

Just now, TheV_Machine said:

Based on These leaked scores, im far from seeing the 1800X beaten in multi threaded workload, but i might Just underestimate the oc scaling of the 8700K

I'm not going based on those since I'm pretty sure they were already debunked.  I'm just taking 1.5x the current 7700k score

[Drak3]

Just now, TahoeDust said:

...and that is only considering multicore performance.  When you start factoring in the huge single core advantage the 8700k has, suddenly Ryzen is not looking so great.

Maybe I should illustrate that point by comparing the 6 core 8700K against the 8 core 1800X, and the 8700K being up its ass in ALL Core performance in the realistic best and worst case for the 8700K against the best option Ryzen has for performance, pyo.

 

Oh, wait, I did, pyo.

 

And I called it out specifically too, pyo.

[TheV_Machine]

1 minute ago, Ryan_Vickers said:

I'm not going based on those since I'm pretty sure they were already debunked

Didnt know about that, do you have a source?

[TahoeDust]

1 minute ago, Drak3 said:

Maybe I should illustrate that point by comparing the 6 core 8700K against the 8 core 1800X, and the 8700K being up its ass in ALL Core performance in the realistic best and worst case for the 8700K against the best option Ryzen has for performance, pyo.

 

Oh, wait, I did, pyo.

 

And I called it out specifically too, pyo.

WTF is pyo?

[Drak3]

Just now, TahoeDust said:

WTF is pyo?

Pyo is pyo, pyo.

[vanished]

Just now, TheV_Machine said:

Didnt know about that, do you have a source?

There was a news post that got moved to the CPUs section about it where the first post was a twitter post showing the numbers were inaccurate.  I don't think I'll be able to find it, but the point is a) there was that twitter post, and b) upon doing the math you'd find that the distance between the multi-threaded and single-threaded score is too small to be correct... it was what you'd see on an i5