[Geekbench] NEW ZEN BENCHMARK!!! - Page 3 - Overclock.net

Forum Jump: 
Reply
 
Thread Tools
post #21 of 54 Old 10-01-2016, 02:21 AM
Overclocker
 
Join Date: Aug 2011
Location: holland
Posts: 113
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 1
sliflex is offline  
Sponsored Links
Advertisement
 
post #22 of 54 Old 10-01-2016, 05:20 AM
x264 Enthusiast
 
Undervolter's Avatar
 
Join Date: Mar 2013
Posts: 4,513
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 514
Quote:
Originally Posted by sliflex View Post

or maybe amd just sucks tongue.gif

That geekbench score, doesn't compute with the video where 3Ghz Zen sligltly beats 3Ghz Broadwell-E. Let's go back at elementary school. "Those who are equal amongst them, are equal amongst third parties".

In AMD's presentation, 8 core Zen is let's say sligthly inferior to 6900K (3.2Ghz). Let's pretend for easyness' sake that they are equal. Blender is fully multithreaded AFAIK.

In that case, we take another fully multithreaded test with known values as guide. Source is CPU Boss:

6900K Cinebench 10 : 48835 (single core: 7618).
FX8350 Cinebench 10: 22674 (single core: 4338).

So Zen 8 core at 3Ghz, should approximately also score 48835, if it's so close to 6900K in Blender. Which means that it should also beat FX heavily.

Which is completely incompatible with Geekbench Naples results for 1.4Ghz.

So, either there is something wrong with Geekbench or with AMD's presentation. It's more probable the former.
jsc1973 likes this.

PLEASE, don't PM me to solve your computer problems privately. That's what the forum is for.
Undervolter is offline  
post #23 of 54 Old 10-01-2016, 12:15 PM
AMD Overclocker
 
BulletBait's Avatar
 
Join Date: Jun 2015
Location: St. Paul, MN
Posts: 1,393
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 150
@Undervolter logic? Blasphemer! We all know that AMD isn't worth the silicon it's built on. You shall repent! Buy 5 new Intel procs as indulgence, or make 30 shill posts as penance, preferably in the AMD subforum for maximum reductions of your sins against Intel.

Otherwise, may your blueless soul rot forever in processor purgatory, sir.
Schmuckley likes this.

BulletBait is offline  
Sponsored Links
Advertisement
 
post #24 of 54 Old 10-01-2016, 12:15 PM
1.7ghz
 
formula m's Avatar
 
Join Date: Jun 2011
Posts: 2,030
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 80
Quote:
Originally Posted by Undervolter View Post


That geekbench score, doesn't compute with the video where 3Ghz Zen sligltly beats 3Ghz Broadwell-E. Let's go back at elementary school. "Those who are equal amongst them, are equal amongst third parties".

In AMD's presentation, 8 core Zen is let's say sligthly inferior to 6900K (3.2Ghz). Let's pretend for easyness' sake that they are equal. Blender is fully multithreaded AFAIK.

In that case, we take another fully multithreaded test with known values as guide. Source is CPU Boss:

6900K Cinebench 10 : 48835 (single core: 7618).
FX8350 Cinebench 10: 22674 (single core: 4338).

So Zen 8 core at 3Ghz, should approximately also score 48835, if it's so close to 6900K in Blender. Which means that it should also beat FX heavily.

Which is completely incompatible with Geekbench Naples results for 1.4Ghz.

So, either there is something wrong with Geekbench or with AMD's presentation. It's more probable the former.

 

 

Say what...?

 

You can't cross compare 2 different benchmarks, on how they scale against each other. It is meaningless.

formula m is offline  
post #25 of 54 Old 10-01-2016, 12:32 PM
x264 Enthusiast
 
Undervolter's Avatar
 
Join Date: Mar 2013
Posts: 4,513
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 514
Quote:
Originally Posted by formula m View Post


Say what...?

You can't cross compare 2 different benchmarks, on how they scale against each other. It is meaningless.

I wasn't trying to compare the benchmarks. I was trying to compare the CPUs using benchmarks that i could find, because i couldn't find in CPU Boss the geekbench 3 score of 6900K, so i had to resort to Cinebench.

Multithreaded applications that load 100% the CPU, have a habbit of showing roughly the same results in all other multithreaded applications. I don't mean as absolute score, but as a relative performance. The main difference being, integer vs FPU use. But in any case, they approximately all give the same image at the end, as % difference. So, if Zen in something that loads 100% 8 cores, is slightly inferior to 6900K, then the same should happen to any other application that can load 100% the CPU. And i don't think AMD would dare make a public demo lying about the frequencies of the 2 CPUs and Blender from what i read, spontanesously uses 100% load. So it can't be "faked", unless AMD lied about the clocks.

The performance of the Zen in Geekbench on the other hand, doesn't arrive not even approximately to what one would expect. Which doesn't make sense at all and indicates that either something is wrong with Geekbench working with a non released CPU or with the person that ran the geekbench and did something wrong.

And there is also something else. You can't have crap single threaded performance and magically achieve miraculous multithreaded performance. I mean, you can't have for example single threaded performance of a 4Ghz Vishera and have multithreaded performance of 6900K. It's just not possible.

To put it otherwise. A 6900K, scores double the score of the FX8350.

Historically, all multithreaded tests, are within 15-20% range from each other, depending on whether they are integer or FPU. But in each category, this is the variation range. See here the multithreaded section:

http://www.cpu-world.com/Compare/436/AMD_FX-Series_FX-8350_vs_Intel_Core_i5_i5-2500K.html


So, if Zen slightly loses by the 6900K, which on its turn scores double the score of a 8350, then Zen can't possibly be as bad as Geekbench says. At least i can't figure out a way of how this can happen. Unless AMD rigged the Blender test.

PLEASE, don't PM me to solve your computer problems privately. That's what the forum is for.
Undervolter is offline  
post #26 of 54 Old 10-01-2016, 12:40 PM
x264 Enthusiast
 
Undervolter's Avatar
 
Join Date: Mar 2013
Posts: 4,513
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 514
Quote:
Originally Posted by BulletBait View Post

@Undervolter logic? Blasphemer! We all know that AMD isn't worth the silicon it's built on. You shall repent! Buy 5 new Intel procs as indulgence, or make 30 shill posts as penance, preferably in the AMD subforum for maximum reductions of your sins against Intel.

Otherwise, may your blueless soul rot forever in processor purgatory, sir.

I confess my sin and i deserve the penance. I will even wear a T-shirt saying "I love Intel" for a week. May the blue God have mercy on my soul! biggrin.gif

PLEASE, don't PM me to solve your computer problems privately. That's what the forum is for.
Undervolter is offline  
post #27 of 54 Old 10-01-2016, 12:44 PM
x264 Enthusiast
 
Undervolter's Avatar
 
Join Date: Mar 2013
Posts: 4,513
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 514
For the record, if i believe this that i just googled, Blender is mainly FPU test:
Quote:
he really important thing for rendering is the FPU (Floating Point Unit), which is already 80bit internal and wont get upgraded.

https://www.blender.org/forum/viewtopic.php?t=4476

Cinebench AFAIK is also FPU heavy. So i don't see how the same CPU can score similarly to 6900K in one 100% FPU heavy load and fail miserably in another 100% FPU heavy load (be it Geekbench or Cinebench).

PLEASE, don't PM me to solve your computer problems privately. That's what the forum is for.
Undervolter is offline  
post #28 of 54 Old 10-01-2016, 02:18 PM
x264 Enthusiast
 
Undervolter's Avatar
 
Join Date: Mar 2013
Posts: 4,513
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 514
GeekBench miracles: How to boost your Zen SHA1 score from 49871 to 148139. rolleyes.gif




http://browser.primatelabs.com/geekbench3/compare/8076870?baseline=8076878

Yes, very reliable.Someone must be having fun, that's for sure.

PLEASE, don't PM me to solve your computer problems privately. That's what the forum is for.
Undervolter is offline  
post #29 of 54 Old 10-01-2016, 02:43 PM
New to Overclock.net
 
kd5151's Avatar
 
Join Date: Jun 2011
Posts: 2,242
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 6 Post(s)
Liked: 131
Shot counter:1
Marios145 likes this.
kd5151 is online now  
post #30 of 54 Old 10-01-2016, 02:47 PM
 
The Robot's Avatar
 
Join Date: Mar 2013
Posts: 1,905
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 124
The Robot is offline  
Reply

Quick Reply

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off