Go Back   Hardware Canucks > HARDWARE > CPU's and Motherboards

    
Reply
 
LinkBack Thread Tools Display Modes
  #61 (permalink)  
Old March 3, 2017, 07:58 AM
Groove's Avatar
MVP
 
Join Date: Jan 2009
Location: Ottawa(ish)
Posts: 387

My System Specs

Default

Quote:
Originally Posted by SKYMTL View Post
I'll keep parroting this again and again: its better to get a native quad core, non-HT / SMT processor for gaming. People seem to be finally waking up to that fact.
Isn't it the lower 99th percentile typically higher with more threads though? My understanding was that with everything else equal, frametime is more representative of gameplay smoothness that raw fps output?
Reply With Quote
  #62 (permalink)  
Old March 3, 2017, 08:12 AM
SKYMTL's Avatar
HardwareCanuck Review Editor
 
Join Date: Feb 2007
Location: Montreal
Posts: 13,605
Default

Quote:
Originally Posted by Groove View Post
Isn't it the lower 99th percentile typically higher with more threads though? My understanding was that with everything else equal, frametime is more representative of gameplay smoothness that raw fps output?
Not necessarily. First of all remember, 99th percentile frame times will also affect averages. As for more threads positively affecting those outcomes, that's a red herring since it varies wildly from one game to another and is often optimization-dependent. Some games (DOOM w/ Vulkan) have great back-ends that schedule tasks in a scalable way across multiple cores with ease. Others (DX12 in BF1) do it really well across 2-4 cores but don't benefit above that.
__________________
Reply With Quote
  #63 (permalink)  
Old March 3, 2017, 08:22 AM
Groove's Avatar
MVP
 
Join Date: Jan 2009
Location: Ottawa(ish)
Posts: 387

My System Specs

Default

Quote:
Originally Posted by SKYMTL View Post
Not necessarily. First of all remember, 99th percentile frame times will also affect averages. As for more threads positively affecting those outcomes, that's a red herring since it varies wildly from one game to another and is often optimization-dependent. Some games (DOOM w/ Vulkan) have great back-ends that schedule tasks in a scalable way across multiple cores with ease. Others (DX12 in BF1) do it really well across 2-4 cores but don't benefit above that.
Definitely makes sense that it's optimization-dependent. I remember seeing something by digital foundry a while back where they showed the 6700K having more stable frametimes than the 6600K in some situations but it only was for 0.1 % of the lows though... so yeah not really a big deal I guess.
Reply With Quote
  #64 (permalink)  
Old March 3, 2017, 08:24 AM
Sagath's Avatar
Moderator
F@H
 
Join Date: Feb 2009
Location: Edmonton, AB
Posts: 3,264

My System Specs

Default

Quote:
Originally Posted by Groove View Post
Isn't it the lower 99th percentile typically higher with more threads though? My understanding was that with everything else equal, frametime is more representative of gameplay smoothness that raw fps output?
This is why people are parroting the 'IPC is worse' fact, regardless of threads. Link back to the HardOCP thread about framerates at 640x480. Processing code faster should theoretically lead to higher minimum frame rates. Well, actually...in the real world that answer isnt that simple and is going to depend on IPC, GPU bottleneck, multi-threading vs multi-cores, and of course, optimizations.

But instead we just tend to focus on the numbers pushed out in an 'apples to apples' comparison to mitigate those factors, which is why some websites include minimum/average/max framerates on their reviews. Of course its not really apples to apples (Because we can make the apple different every time) even then when you look at those numbers, but hey, you've got to start somewhere...

I guess to simplify; any time a review site does these things you can change the perspective of focus, and it COULD cause a different result. Inversely, if you're looking for a specific result it can cause you to bias the methodology.
__________________
My Disclaimer to any advice or comment I make;
Quote:
Originally Posted by CroSsFiRe2009 View Post
I'm a self certified whizbang repair technician with 20 years of professional bullshit so I don't know what I'm talking about
Reply With Quote
  #65 (permalink)  
Old March 3, 2017, 12:19 PM
Master_Shake_'s Avatar
Top Prospect
 
Join Date: May 2011
Location: Ontario, Canada
Posts: 183

My System Specs

Default

Skymtl,

why were the details changed for all the benchmarks compared to previous ones?

was it the addition of a newer gpu?
Reply With Quote
  #66 (permalink)  
Old March 3, 2017, 12:57 PM
SKYMTL's Avatar
HardwareCanuck Review Editor
 
Join Date: Feb 2007
Location: Montreal
Posts: 13,605
Default

Correct. I was using a GTX 980 or GTX 1070 before which caused SIGNIFICANT top-end bottlenecking in games, especially DX12.
__________________
Reply With Quote
  #67 (permalink)  
Old March 3, 2017, 01:01 PM
Master_Shake_'s Avatar
Top Prospect
 
Join Date: May 2011
Location: Ontario, Canada
Posts: 183

My System Specs

Default

Quote:
Originally Posted by SKYMTL View Post
Correct. I was using a GTX 980 or GTX 1070 before which caused SIGNIFICANT top-end bottlenecking in games, especially DX12.

would you mind posting a bench of your test bench with the same settings as before?

just the GTA V one is all.
Reply With Quote
  #68 (permalink)  
Old March 3, 2017, 01:41 PM
SKYMTL's Avatar
HardwareCanuck Review Editor
 
Join Date: Feb 2007
Location: Montreal
Posts: 13,605
Default

Quote:
Originally Posted by Master_Shake_ View Post
would you mind posting a bench of your test bench with the same settings as before?

just the GTA V one is all.
I would, but the 980 has been retired to my storage locker across town.
__________________
Reply With Quote
  #69 (permalink)  
Old March 3, 2017, 05:19 PM
Fleurious's Avatar
Top Prospect
F@H
 
Join Date: Jun 2011
Location: Tottenham, ON
Posts: 167

My System Specs

Default

Can't wait to see how the rest of the lineup performs. I've been aching to replace the Q9550 with something a bit more modern and ideally micro/mini ATX.
Reply With Quote
  #70 (permalink)  
Old March 3, 2017, 08:51 PM
MVP
 
Join Date: Dec 2012
Location: Chilliwack...Time Is Just A Rubber Band....
Posts: 383
Default

The Ryzen 7 1800X has dethroned the former king of 8C/16T, the I7 5960X in Cinebench R11.5/R15 and GPUPI for CPU 1B and done it at significantly lower clocks. Check it out at Launch Day! AMD Ryzen 7: 3 Global First Places.
Reply With Quote
Reply


Thread Tools
Display Modes

Similar Threads
Thread Thread Starter Forum Replies Last Post
The AMD Ryzen CPU Preview; Zen Matures - Comment Thread SKYMTL CPU's and Motherboards 23 December 27, 2016 08:54 PM
The AMD R9 390X 8GB Performance Review Comment Thread SKYMTL Video Cards 37 July 9, 2015 03:17 PM
NVIDIA GTX 970 SLI Performance Review (Comment Thread) SKYMTL Video Cards 12 November 18, 2014 05:31 PM
Far Cry 2 Hardware Performance Review Comment Thread FiXT Gaming 21 November 30, 2008 09:09 PM