Hardware Secrets Forums


Go Back   Hardware Secrets Forums > Miscellaneous > Content Comments



Reply
 
Thread Tools Search this Thread Display Modes
Old 11-03-2011, 04:06 PM   #21
klinc
Senior Member
 
klinc's Avatar
 
Join Date Dec 2010
Posts: 323
klinc is on a distinguished road

Default

Quote:
Originally Posted by Gabriel Torres View Post
You guys have to make up your mind. If I increase details, people complain that we are not stressing the CPU, but the video card. If we lower them, then come people like you complaining as well... I believe there is no way to create the perfect methology, there will always be people complaining about our methodology, no matter how we decide to run our tests...
Correct How do you really test a cpu at high resolutions? Artificially create a bottleneck?
Gabriel have anyone tried Warp where you can play modern games in windows without a decent gpu? Have anyone tried benching a game on that? Would be interesting to see
http://msdn.microsoft.com/en-us/libr...=vs.85%29.aspx.

Quote:
Originally Posted by Cheetos View Post
What's the point in test a high end CPU with low resolutions ?

That's not reality, that's a useless test.
Low resolutions the gpu will run faster making it harder for the cpu to catch up. I guess the people want to see if it will bottleneck at stock with a high end gpu. But why must you guys test for people who decided to overkill and not buying according to their monitors but their egos
__________________
I7-2600k @4.8Ghz | ASUS MAXIMUS IV EXTREME 3-Way SLI | 2x4gb Corsair Vengeance 1867mhz CL8 10-9-24 1T 1.600V | 3x Zotac Gtx580 SLI | OCZ Vertex 2 SF 180gb | 2x Samsung F3 Spinpoint 1Tb Raid-0 | CORSAIR AX1200 | Thermalright Silver Arrow | Antec 1200 |
3DMARK11
klinc is offline   Reply With Quote
new Sponsored Links

Old 11-03-2011, 04:11 PM   #22
Cheetos
Senior Member
 
Cheetos's Avatar
 
Join Date Jun 2008
Brazil
Posts: 426
Cheetos is on a distinguished road

Default

Quote:
Originally Posted by klinc View Post
Correct How do you really test a cpu at high resolutions? Artificially create a bottleneck?
How can you say that a CPU is better than other in games using low resolutions with a high end videocard ?
That's not reality.

Quote:
Originally Posted by klinc View Post
Low resolutions the gpu will run faster making it harder for the cpu to catch up.
I know that.

Quote:
Originally Posted by klinc View Post
I guess the people want to see if it will bottleneck at stock with a high end gpu.
The test precisely shows this.

Quote:
Originally Posted by klinc View Post
But why must you guys test for people who decided to overkill and not buying according to their monitors but their egos
??
__________________
Goodbye, and thanks for all the fish !
Cheetos is offline   Reply With Quote
Old 11-03-2011, 04:17 PM   #23
klinc
Senior Member
 
klinc's Avatar
 
Join Date Dec 2010
Posts: 323
klinc is on a distinguished road

Default

Quote:
Originally Posted by Cheetos View Post
How can you say that a CPU is better than other in games using low resolutions with a high end videocard ?
That's not reality.
No im not saying that. With these cpu tests at higher resolutions I really don't pay much attention to it. Because at 1080p using a X4 or my 2600k im going to get the same results using the same gpu. Theres no better gaming performance thats where the gpu matters unless you play a extremely bad port. (Bring back GTAIV for some of the tests.) But warp doesn't need a gpu its all cpu. We are talking about testing the cpu right
Just have a look here
http://msdn.microsoft.com/en-us/libr...=vs.85%29.aspx

Quote:
Originally Posted by Cheetos View Post
How can you say that a CPU is better than other in games using low resolutions with a high end videocard ?
That's not reality.



I know that.



The test precisely shows this.



??
Thats what a test at 1280 will show. Running a 580 or a 6970 at that resolution its going to get through its workload like theres no tomorrow where the cpu will struggle to keep up. But do it at the gpus recommended resolution and the cpu wont have such a hard time to keep up because the gpu got more of a workload thats slows it down.
See what a overkill can do?
__________________
I7-2600k @4.8Ghz | ASUS MAXIMUS IV EXTREME 3-Way SLI | 2x4gb Corsair Vengeance 1867mhz CL8 10-9-24 1T 1.600V | 3x Zotac Gtx580 SLI | OCZ Vertex 2 SF 180gb | 2x Samsung F3 Spinpoint 1Tb Raid-0 | CORSAIR AX1200 | Thermalright Silver Arrow | Antec 1200 |
3DMARK11
klinc is offline   Reply With Quote
Old 11-03-2011, 04:25 PM   #24
Cheetos
Senior Member
 
Cheetos's Avatar
 
Join Date Jun 2008
Brazil
Posts: 426
Cheetos is on a distinguished road

Default

Quote:
Originally Posted by klinc View Post
Thats what a test at 1280 will show
No, it doesn't show anything, it's a useless test.

Instead of suggesting that, the use of "heavy CPU load games" is a very good idea.

Despite GTA4, I don't know wich games could be tested, so, I'll pass.
__________________
Goodbye, and thanks for all the fish !
Cheetos is offline   Reply With Quote
Old 11-03-2011, 05:05 PM   #25
trek554
Junior Member
 
Join Date Nov 2011
Posts: 5
trek554 is on a distinguished road

Default

Quote:
Originally Posted by Cheetos View Post
What's the point in test a high end CPU with low resolutions ?

That's not reality, that's a useless test.
wow, are you freaking joking? you can use a low res test to see how much actual cpu power is there. if one cpu is way faster than the other then you know that cpu will be stronger in cpu limited situations and will not hold back those future gpu upgrades.

if all you do is a high res test then that is nothing but a gpu benchmark. of course all high end cpus will be bunched right together with a single gpu at 1920x1080.

and AGAIN running low settings is useless because MANY settings impact the cpu. I have already told you the best way to test cpus is at 1280x720 with high settings and no AA. and then run those same high settings and no AA at 1920x1080 for those that want to see it at realistic conditions.

that will cover basically every situation and make everybody happy.

Last edited by trek554; 11-03-2011 at 05:07 PM.
trek554 is offline   Reply With Quote
Old 11-03-2011, 05:15 PM   #26
klinc
Senior Member
 
klinc's Avatar
 
Join Date Dec 2010
Posts: 323
klinc is on a distinguished road

Default

Quote:
Originally Posted by Cheetos View Post
No, it doesn't show anything, it's a useless test.

Instead of suggesting that, the use of "heavy CPU load games" is a very good idea.

Despite GTA4, I don't know wich games could be tested, so, I'll pass.
Thats easy. The bad ports.
GTA IV
Black Ops
Fallout 3 and New Vegas
DMC3
Bulletstorm

theres lots of them or you can test WARP.

Quote:
Originally Posted by trek554 View Post
wow, are you freaking joking? you can use a low res test to see how much actual cpu power is there. if one cpu is way faster than the other then you know that cpu will be stronger in cpu limited situations and will not hold back those future gpu upgrades.

if all you do is a high res test then that is nothing but a gpu benchmark. of course all high end cpus will be bunched right together with a single gpu at 1920x1080.

and AGAIN running low settings is useless because MANY settings impact the cpu. I have already told you the best way to test cpus is at 1280x720 with high settings and no AA. and then run those same high settings and no AA at 1920x1080 for those that want to see it at realistic conditions.

that will cover basically every situation and make everybody happy.
that will have more of a chance bottlenecking the cpu than on a higher resolution. On 1280 its the resolution where the cpu still have a influence. Agree
__________________
I7-2600k @4.8Ghz | ASUS MAXIMUS IV EXTREME 3-Way SLI | 2x4gb Corsair Vengeance 1867mhz CL8 10-9-24 1T 1.600V | 3x Zotac Gtx580 SLI | OCZ Vertex 2 SF 180gb | 2x Samsung F3 Spinpoint 1Tb Raid-0 | CORSAIR AX1200 | Thermalright Silver Arrow | Antec 1200 |
3DMARK11
klinc is offline   Reply With Quote
Old 11-03-2011, 08:58 PM   #27
trek554
Junior Member
 
Join Date Nov 2011
Posts: 5
trek554 is on a distinguished road

Default

remember how I said if you run all high setting and use a low res with a single gpu that it would give you an idea of how the cpu will handle being cpu limited or future gpu upgrades. here you can see just how Bulldozer stacks up when pushing 2 and 3 gpus even at very high res. once the 3rd gpu is added, Bulldozer becomes a joke in most cases. heck in some cases it did not even look worth a poo with just 2 cards and thats a whopping 5040x1050 res in most tests. not to mention the horrible inconsistent framerates and very low minimums the 8150 showed in most cases. well many people at just 1920x1080 and looking at Bulldozer might be thinking about sli/crossfire and or upgrading to much faster gpus at some time. just imagine how disappointed they would be.


http://www.hardocp.com/article/2011/...mance_review/2


All of our gameplay testing has brought us to this single conclusion, if you are serious about gaming on the PC, and or you have a dual-GPU or better configuration, you will simply have a better gameplay experience with an Intel i5 or i7 CPU. Currently, when it comes to Bulldozer and Sandy Bridge, Intel simply has the better platform for high-end gaming.

Last edited by trek554; 11-03-2011 at 10:12 PM.
trek554 is offline   Reply With Quote
Old 11-04-2011, 06:41 AM   #28
klinc
Senior Member
 
klinc's Avatar
 
Join Date Dec 2010
Posts: 323
klinc is on a distinguished road

Default

Its because everyone went and benched it core vs core Thx to Amd and their naming scheme. Its 4 modules / 8 threads not 8 cores. Do you benchmark Intels hypertreading Cores against real ones?
__________________
I7-2600k @4.8Ghz | ASUS MAXIMUS IV EXTREME 3-Way SLI | 2x4gb Corsair Vengeance 1867mhz CL8 10-9-24 1T 1.600V | 3x Zotac Gtx580 SLI | OCZ Vertex 2 SF 180gb | 2x Samsung F3 Spinpoint 1Tb Raid-0 | CORSAIR AX1200 | Thermalright Silver Arrow | Antec 1200 |
3DMARK11
klinc is offline   Reply With Quote
Old 08-17-2012, 11:42 PM   #29
adlihajarat
Junior Member
 
Join Date Aug 2012
Posts: 1
adlihajarat is on a distinguished road

Default AMD VS Intel core wars

Has it occurred to anyone that the Big Endian vs Little Endian plays a role in this battle, and that the way the test bench software is compiled plays a big role.

I once wrote an email to a ZD.net evaluator and demonstrated him, that the way the software optimization is used affects drastically the speed of program execution.
For example, write a simple C program that does matrix multiplication using an unbiased compiler such as blood shed and optimize the code twice on for intel micro code and once for AMD microcode and benchmark the same program on one intel and one AMD platforms and see the difference.
You would notice that the one optimized for intel runs much faster that the other code if run on intel platform, and vice versa.
That is all due to the different ways intel and AMD process their data.
Hope this gives you an insight for future comparisons.

Last edited by adlihajarat; 08-17-2012 at 11:48 PM. Reason: misspelling
adlihajarat is offline   Reply With Quote
new Sponsored Links

Reply

Share This Thread

Bookmarks

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Forum Jump


All times are GMT -8. The time now is 09:05 PM.


vBulletin Security provided by vBSecurity (Lite) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
vBulletin Optimisation provided by vB Optimise (Lite) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd. ()
2004-12, Hardware Secrets, LLC. All rights reserved.