EVGA

Helpful ReplyNew benchmark: NVBench!

Page: 1234 > Showing page 1 of 4
Author
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
2011/10/29 17:24:38 (permalink)
 
Let's see your scores!
This program is also good as a stability test.
Watch out, Heavy can take a few hours on two gtx 580 in SLI.
"Normal" takes 6 minutes on my system.
 
Have fun!
 
Note: this program will use all available cards no matter if SLI is enabled or not.
If you like a new feature, write it in the comments.
 
Leaderboard
 
post edited by Zibri - 2011/10/31 19:09:05
#1
MajestacaL
FTW Member
  • Total Posts : 1314
  • Reward points : 0
  • Joined: 2006/06/22 19:55:11
  • Status: offline
  • Ribbons : 6
Re:New benchmark: NVBench! 2011/10/29 17:48:44 (permalink)
Here is mine. Nice work on the benchmark thanks!
 


 
EVGA Affiliate Code: QFVZM37BXK
STEAM Profile http://steamcommunity.com/id/majestacal
EA ORIGIN ID: MajestacaL
#2
shaneduce
CLASSIFIED Member
  • Total Posts : 2038
  • Reward points : 0
  • Joined: 2009/10/21 10:19:14
  • Location: Portland ,Or.
  • Status: offline
  • Ribbons : 3
Re:New benchmark: NVBench! 2011/10/29 17:57:54 (permalink)
Ty will dl and bench tonight when I get home.
#3
Little Sister
CLASSIFIED ULTRA Member
  • Total Posts : 6076
  • Reward points : 0
  • Joined: 2011/06/06 08:01:11
  • Status: offline
  • Ribbons : 10
Re:New benchmark: NVBench! 2011/10/29 18:29:56 (permalink)


INTEL I7 2600K  -  CORSAIR H100  -  EVGA Z68 FTW  -  CORSAIR DOMINATOR GT 2000MHZ 16GB  -  EVGA GTX 580 3GB SLI X3 
WESTERN DIGITAL BLACK 1TB X2  -  CORSAIR FORCE GT 120GB X2  -  PLEXTOR BLU-RAY BURNER  -  CORSAIR AX1200
ENERMAX FULMO GT  -  WINDOWS 7 ULTIMATE X64 

#4
EVGAUser0319
SSC Member
  • Total Posts : 620
  • Reward points : 0
  • Joined: 2007/08/28 21:31:50
  • Status: offline
  • Ribbons : 1
Re:New benchmark: NVBench! 2011/10/29 18:33:18 (permalink)

 
While watching the Flyers game on Windows 7 Media Center, so its probably hurt a bit.  Ran in like 2 minutes.

EVGA x58 E760 Classified
i7 920 @ 4.3 GHz
Tri-SLI EVGA GTX 580 SC
OCZ Gold 12GB DDR3 12800
Thermaltake 1200W PSU
LG W3000H 30" 2560x1600
#5
EVGAUser0319
SSC Member
  • Total Posts : 620
  • Reward points : 0
  • Joined: 2007/08/28 21:31:50
  • Status: offline
  • Ribbons : 1
Re:New benchmark: NVBench! 2011/10/29 18:35:17 (permalink)

 
Media center closed

EVGA x58 E760 Classified
i7 920 @ 4.3 GHz
Tri-SLI EVGA GTX 580 SC
OCZ Gold 12GB DDR3 12800
Thermaltake 1200W PSU
LG W3000H 30" 2560x1600
#6
owentheomen
Superclocked Member
  • Total Posts : 183
  • Reward points : 0
  • Joined: 2011/05/01 19:58:42
  • Location: Reno, Nevada
  • Status: offline
  • Ribbons : 0
Re:New benchmark: NVBench! 2011/10/29 19:24:55 (permalink)
works pretty good :D



#7
Phillsta
iCX Member
  • Total Posts : 346
  • Reward points : 0
  • Joined: 2011/09/21 01:18:35
  • Status: offline
  • Ribbons : 0
Re:New benchmark: NVBench! 2011/10/30 01:22:17 (permalink)
3x Gtx570,s @ 860/1720/2075 they get 55 @stock

Attached Image(s)


<<-- Asus Maximus IV Extreme Z --><--Intel Core i7 2700K--><--3xAsus 7970--><--8g Corsair Dominator GT @2000 9-9-9-24--><--Silverstone ST1500 Psu--><--Xigmatek Elisium--><--Mobo Block By EK--><--Cpu block EK Supreme HF--><--Gpu Blocks by EK--><--Black Ice 360+240 rads--><--2x Koolance 450-S Pumps--><--Koolance Twin Res/Pump Enclosure--><--3Dmark11http://hwbot.org/submission/2266705_                                                                                    3Dmark Vantagehttp://hwbot.org/submission/2266708_ 
#8
HeavyHemi
Omnipotent Enthusiast
  • Total Posts : 14099
  • Reward points : 0
  • Joined: 2008/11/28 20:31:42
  • Location: Western Washington
  • Status: online
  • Ribbons : 96
Re:New benchmark: NVBench! 2011/10/30 01:52:59 (permalink)
Doesn't seem to recognize SLI
 

Attached Image(s)


EVGA X99 FTWK / i7 6850K 4.5ghz / GTX1080 Ti FE Hybrid / 32GB Corsair LPX 3200mhz / Samsung 850Pro 256GB / Corsair AX1200 / Window 10 Pro
Fire Strike                        24,163
Fire  Strike Extreme        14,452 
Fire Strike Ultra                7,711 
Time Spy                        10,357

#9
Phillsta
iCX Member
  • Total Posts : 346
  • Reward points : 0
  • Joined: 2011/09/21 01:18:35
  • Status: offline
  • Ribbons : 0
Re:New benchmark: NVBench! 2011/10/30 02:02:44 (permalink)
what makes you say that?

<<-- Asus Maximus IV Extreme Z --><--Intel Core i7 2700K--><--3xAsus 7970--><--8g Corsair Dominator GT @2000 9-9-9-24--><--Silverstone ST1500 Psu--><--Xigmatek Elisium--><--Mobo Block By EK--><--Cpu block EK Supreme HF--><--Gpu Blocks by EK--><--Black Ice 360+240 rads--><--2x Koolance 450-S Pumps--><--Koolance Twin Res/Pump Enclosure--><--3Dmark11http://hwbot.org/submission/2266705_                                                                                    3Dmark Vantagehttp://hwbot.org/submission/2266708_ 
#10
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 05:48:52 (permalink)
@HeavyHemi: yes it does and it uses it. Even with SLI disabled, this will use ALL available cards.
 
@Phillsta: please re-run and post the link to the result not the image. Valid scores are the ones online only.
#11
Krony
FTW Member
  • Total Posts : 1192
  • Reward points : 0
  • Joined: 2010/11/10 22:42:31
  • Location: England
  • Status: offline
  • Ribbons : 6
Re:New benchmark: NVBench! 2011/10/30 06:44:44 (permalink)
Sli 3Gb 580's @ 954/2025

Link
post edited by Krony - 2011/10/30 07:02:20


#12
CraptacularOne
Omnipotent Enthusiast
  • Total Posts : 12562
  • Reward points : 0
  • Joined: 2006/06/12 17:20:44
  • Location: Florida
  • Status: offline
  • Ribbons : 173
Re:New benchmark: NVBench! 2011/10/30 07:00:12 (permalink)

 
What exactly is this measuring? I have 2 GTX480s @ 800/1600/4000 and a GTX460 SE @ 850/1700/1800. A load was placed on all of them during the test. How is this app using 2 different GPUs?

i9 7960X @ 4.6Ghz H2O // RTX 2080Ti FE H2O // Viotek SUW49C 32:9 Ultra Wide // ASUS Prime X299 Deluxe // 32GB G.Skill Trident Z RGB 3600Mhz // EVGA G3 SuperNova 1000w 
 
Ryzen 5 3600 @ 4.2Ghz // RTX 2070 // Oculus Rift S // ASUS ROG Strix B450i // 16GB Team Group 3600Mhz // EVGA G3 SuperNova 750w
#13
Krony
FTW Member
  • Total Posts : 1192
  • Reward points : 0
  • Joined: 2010/11/10 22:42:31
  • Location: England
  • Status: offline
  • Ribbons : 6
Re:New benchmark: NVBench! 2011/10/30 07:04:47 (permalink)
It must use all gpu's somehow as ur score is the same as mine and sli 480's @ 800/2000 would normally score alot less than sli 580's @ 954/2025 so that 460 is really helping in this bench, might have to dig out my 9800GTX+ :)


#14
CraptacularOne
Omnipotent Enthusiast
  • Total Posts : 12562
  • Reward points : 0
  • Joined: 2006/06/12 17:20:44
  • Location: Florida
  • Status: offline
  • Ribbons : 173
Re:New benchmark: NVBench! 2011/10/30 07:09:02 (permalink)
Disabling Surround brings up the score a few points:
 

 
Not really sold on this being an accurate benchmark yet. I would really like to know what it is measuring.

i9 7960X @ 4.6Ghz H2O // RTX 2080Ti FE H2O // Viotek SUW49C 32:9 Ultra Wide // ASUS Prime X299 Deluxe // 32GB G.Skill Trident Z RGB 3600Mhz // EVGA G3 SuperNova 1000w 
 
Ryzen 5 3600 @ 4.2Ghz // RTX 2070 // Oculus Rift S // ASUS ROG Strix B450i // 16GB Team Group 3600Mhz // EVGA G3 SuperNova 750w
#15
CraptacularOne
Omnipotent Enthusiast
  • Total Posts : 12562
  • Reward points : 0
  • Joined: 2006/06/12 17:20:44
  • Location: Florida
  • Status: offline
  • Ribbons : 173
Re:New benchmark: NVBench! 2011/10/30 07:16:22 (permalink)
Krony

It must use all gpu's somehow as ur score is the same as mine and sli 480's @ 800/2000 would normally score alot less than sli 580's @ 954/2025 so that 460 is really helping in this bench, might have to dig out my 9800GTX+ :)

If it was using all 3 cards, then my score should be higher than yours. Between your 2 cards you have 1024 cores running at 954Mhz while I have collectively 1248 cores between my 3 cards running at 800-850Mhz. If this was a CUDA benchmark or a collective performance of 3 cards combined then my score should be higher by a good margin.
post edited by CraptacularOne - 2011/10/30 07:18:53

i9 7960X @ 4.6Ghz H2O // RTX 2080Ti FE H2O // Viotek SUW49C 32:9 Ultra Wide // ASUS Prime X299 Deluxe // 32GB G.Skill Trident Z RGB 3600Mhz // EVGA G3 SuperNova 1000w 
 
Ryzen 5 3600 @ 4.2Ghz // RTX 2070 // Oculus Rift S // ASUS ROG Strix B450i // 16GB Team Group 3600Mhz // EVGA G3 SuperNova 750w
#16
Krony
FTW Member
  • Total Posts : 1192
  • Reward points : 0
  • Joined: 2010/11/10 22:42:31
  • Location: England
  • Status: offline
  • Ribbons : 6
Re:New benchmark: NVBench! 2011/10/30 07:18:41 (permalink)
Yea it's strange really how it works.
Edit: Don't think i can get my 9800GTX+ in as the fitting and tubing on the bottom card is gonna be in the way, bottom card will be running @ 4x anyway.


#17
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 07:24:33 (permalink)
I'm thinking of giving you another digit just for accuracy.
 
For example: my cards get 50 @ 925 and 52 @ 980...
Maybe that would read better as 50.3 and 52.8  or 503 and 528.
What do you think?
I used 2 digits only because the third can fluctuate on the same system.
 
Edit: nevermind, I changed it. 3 digits score now. V1.0 was this score divided by ten and rounded.
 
post edited by Zibri - 2011/10/30 08:07:35
#18
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 07:28:54 (permalink) ☄ Helpful
CraptacularOne

Disabling Surround brings up the score a few points:


Not really sold on this being an accurate benchmark yet. I would really like to know what it is measuring.

This benchmark measures the pure computing power of the cards and as such, everything else is just a matter of settings and optimizations in games, control panel, etc.
I always thought it's useless to measure FPS in a benchmark because FPS can be easily upped or lowered by changing the graphic settings.
 
#19
CraptacularOne
Omnipotent Enthusiast
  • Total Posts : 12562
  • Reward points : 0
  • Joined: 2006/06/12 17:20:44
  • Location: Florida
  • Status: offline
  • Ribbons : 173
Re:New benchmark: NVBench! 2011/10/30 07:29:38 (permalink)
Zibri

I'm thinking of giving you another digit just for accuracy.

For example: my cards get 50 @ 925 and 52 @ 980...
Maybe that would read better as 50.3 and 52.8  or 503 and 528.
What do you think?
I used 2 digits only because the third can fluctuate on the same system.


 
What is the benchmark measuring? Is it just measuring the raw compute power of any GPUs on a system? Also how is it placing a load on 2 different types of GPUs in the same system like mine for example?
 
EDIT: just saw your post above. If it is a collective measurement of GPUs in a system, then I would think it needs some tweaking of the code becasue it doesn't seem to be representing accurate measurement of said GPUs. For instance take my system with it's available 1248 cores of compute power, it should score quite a bit better than a system with only 1024 cores available to it. But if you look at the above results, you'll see that this is not the case.
post edited by CraptacularOne - 2011/10/30 07:32:27

i9 7960X @ 4.6Ghz H2O // RTX 2080Ti FE H2O // Viotek SUW49C 32:9 Ultra Wide // ASUS Prime X299 Deluxe // 32GB G.Skill Trident Z RGB 3600Mhz // EVGA G3 SuperNova 1000w 
 
Ryzen 5 3600 @ 4.2Ghz // RTX 2070 // Oculus Rift S // ASUS ROG Strix B450i // 16GB Team Group 3600Mhz // EVGA G3 SuperNova 750w
#20
maniacvvv
Omnipotent Enthusiast
  • Total Posts : 10053
  • Reward points : 0
  • Joined: 2009/09/10 20:29:17
  • Location: Miami Beach FL
  • Status: offline
  • Ribbons : 164
Re:New benchmark: NVBench! 2011/10/30 07:46:29 (permalink)
As the "author" of this program, please explain how it functions.
 
Please detail what API's are used to place GPU loading across multi GPU setups?
 
Please explain in detail the shader algorithm used and what loading levels it is designed produce?
 
What software steps have you taken to ensure that Nvidia hadware OCP is not bypassed?
 
Please explain in detail, what are the max possible volatge levels when using your program and how far your program will push over the max possible PCIE slot voltage limit.
 
As the person who wrote the program, these are very easy questions to answer, I ask you to do so. 
#21
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 07:59:50 (permalink) ☄ Helpful
I updated the program.
Please re-run it and update the images in your posts.
 
@Maniacvvv: yes they are very easy to answer and I could, but I don't like your attitude and the way you talk to me and to others, so please leave this thread or just post a result.
 
@Everyone else:
This program uses CUDA. No tricks used. So it's perfectly safe as long as cuda implementation is safe :)
 
#22
maniacvvv
Omnipotent Enthusiast
  • Total Posts : 10053
  • Reward points : 0
  • Joined: 2009/09/10 20:29:17
  • Location: Miami Beach FL
  • Status: offline
  • Ribbons : 164
Re:New benchmark: NVBench! 2011/10/30 08:11:51 (permalink)
Zibri



@Maniacvvv: yes they are very easy to answer and I could, but I don't like your attitude and the way you talk to me and to others, so please leave this thread or just post a result.




Given the above statement...  I suggest that users not download or use the program.
 
There is no reason for a programs author, not to answer honest valid questions about it.
 
Given Zibri's previous postings showing a lack of basic Nvidia driver understanding and his repeated novice "mistakes" when making posts concerning Nvidia drivers and settings, and the fact that he has only been using a SLI setup for a few days. I must conclude that if he is unwilling to answer questions about his NVBench program, it is because he is NOT able to do so. 
 
He is welcome to answer the questions I asked, or he should remove the program as using it could be dangerous to users systems and videocards.
 
Answer the questions, or remove the post. 
#23
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 08:13:13 (permalink)
CraptacularOne
What exactly is this measuring? I have 2 GTX480s @ 800/1600/4000 and a GTX460 SE @ 850/1700/1800. A load was placed on all of them during the test. How is this app using 2 different GPUs?

This app will use any GPU present roughly at 99%.
And the score will be the same as the SUM of three single card scores (with the same cards).
 
So if (for example) a stock clocked reference GTX 580 gives about 208 (or 20 in v1.0), two of them will give you about 417 (or 41 in V1.0) and three of them will give you about 625 (or 62 in V1.0)...
 
#24
maniacvvv
Omnipotent Enthusiast
  • Total Posts : 10053
  • Reward points : 0
  • Joined: 2009/09/10 20:29:17
  • Location: Miami Beach FL
  • Status: offline
  • Ribbons : 164
Re:New benchmark: NVBench! 2011/10/30 08:17:39 (permalink)
As the "author" of this program, I ask you AGAIN please explain how it functions.
 
Please detail what API's are used to place GPU loading across multi GPU setups?
 
Please explain in detail the shader algorithm used and what loading levels it is designed produce?
 
What software steps have you taken to ensure that Nvidia hadware OCP is not bypassed?
 
Please explain in detail, what are the max possible volatge levels when using your program and how far your program will push over the max possible PCIE slot voltage limit. 
 
Either answer these questions or I will refer this post to the EVGA Mods
#25
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 08:18:54 (permalink)
maniacvvv
 He is welcome to answer the questions I asked, or he should remove the program as using it could be dangerous to users systems and videocards.
 Answer the questions, or remove the post. 

 
1) I don't take orders from anyone. Who the heck are you?
2) As I already (politely) answered, my program uses CUDA and takes no other actions than measuring the calculation speed.
 
*Removed*
 
post edited by Afterburner - 2011/10/30 08:48:21
#26
Afterburner
EVGA Forum Moderator
  • Total Posts : 25390
  • Reward points : 0
  • Joined: 2007/09/21 14:41:48
  • Location: It's... Classified Yeeeeeeeeeeaaaaaaah........
  • Status: offline
  • Ribbons : 108
Re:New benchmark: NVBench! 2011/10/30 08:26:29 (permalink)
EVGA Mod here.....
 
If the OP does not want to express or answer a question from a member, that IS NOT a TOS violation.
 
Please keep it civil. And not everyone has to agree to agree. Many of us like new things like this and are happy to help. Even if in the end we find out it did nothing for us.
 
Any further repeated request that have already been answered/replied to or asked will be deleted...

#27
billythekid2012
CLASSIFIED Member
  • Total Posts : 2030
  • Reward points : 0
  • Joined: 2009/02/08 17:21:31
  • Status: offline
Re:New benchmark: NVBench! 2011/10/30 08:37:14 (permalink)
running two gtx 480 1700/2000 -2600k @ 4.6 but why is evry one only getting like 50 to 65 in the run.


  

#28
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 08:38:13 (permalink)
I coded a program, and I didn't point a gun at anyone to use it.
I just wanted this thread to be clean so that people could compare results and settings.
 
This benchmark is intended both as a stability test (the Heavy and Infinite modes) both as a simple speed test that won't be affected by the "usual cheats".
 
I wonder why some people just have the urge to spit on everything.
 
post edited by Zibri - 2011/11/10 18:02:44
#29
Zibri
iCX Member
  • Total Posts : 350
  • Reward points : 0
  • Joined: 2011/09/30 09:33:17
  • Location: Italy
  • Status: offline
  • Ribbons : 2
Re:New benchmark: NVBench! 2011/10/30 08:41:35 (permalink)
billythekid2012

running two gtx 480 1700/2000 -2600k @ 4.6 but why is evry one only getting like 50 to 65 in the run.


Because they used V1.0... so you must multiply those scores by 10.
 
#30
Page: 1234 > Showing page 1 of 4
Jump to:
  • Back to Mobile