Tharifas88
New Member
- Total Posts : 5
- Reward points : 0
- Joined: 2011/03/28 12:20:07
- Status: offline
- Ribbons : 0
Hi guys
as the title states I have been having some issues with my setup. I just recently bought two 1080ti for 500 euros each , was a steal really. Some promotion in my country.
Anyhow... I have been having low usage on both cards. I play games at 1440p , I am well aware that most "High" end cards top out @ 4k and nothing less but 1440p should tax them at least more then 70% , right ?
Tested them with sli off and as expected the card/s reach 99%. Sometimes even having better frames.
As for my system specs:
CPU: 4930k
GPU: Title.
HDD: 500GB Samsung SSD
RAM: 16GB 2133Mhzw Corsair Dominator
MOB: RIVE IV
Do you guys think that I am actually starting to hit a cpu bottleneck with my 4930k ? I am seeing my usage that low in most games I have tested: R6S, BF1 those are the games that have the nicest fidelity.
I am thinking about upgrading my cpu/mob and ram. What do you guys think ?
ALSO: Wait for the 9900k or go for the ryzen 2700 ?
Kind Regards
Toby
|
bdary
Omnipotent Enthusiast
- Total Posts : 10331
- Reward points : 0
- Joined: 2008/04/25 14:08:16
- Location: Florida
- Status: offline
- Ribbons : 116
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/21 14:55:25
(permalink)
If a single card is hitting 99%, but then running sli gives you about half that, I'd say it sounds about right. It wouldn't be a 50/50 split between both cards, but you get the idea.
|
Tharifas88
New Member
- Total Posts : 5
- Reward points : 0
- Joined: 2011/03/28 12:20:07
- Status: offline
- Ribbons : 0
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/21 15:22:13
(permalink)
Yeah I thought as much... Well its about time for that upgrade anyhow.... I mean the 4930 was/is a good cpu but it can only go "so" far. I am seeing around 75-80% usage on one card and 55% on the other when SLI is active. So indeed your 50/50 split comment makes sense. Sooo am asking again "just in case you have got some insight in that manner" what cpu should I get ? I mean the 9700k and 9900k are around the corner if you believe the rumor mill. Might be getting the 9900k IF the price isnt completely off this planet like the damn 2080ti.... Thanks mate.
post edited by Sajin - 2018/09/21 15:42:55
|
HeavyHemi
Insert Custom Title Here
- Total Posts : 15665
- Reward points : 0
- Joined: 2008/11/28 20:31:42
- Location: Western Washington
- Status: offline
- Ribbons : 135
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/21 15:43:51
(permalink)
Tharifas88
Hi guys
as the title states I have been having some issues with my setup. I just recently bought two 1080ti for 500 euros each , was a steal really. Some promotion in my country.
Anyhow... I have been having low usage on both cards. I play games at 1440p , I am well aware that most "High" end cards top out @ 4k and nothing less but 1440p should tax them at least more then 70% , right ?
Tested them with sli off and as expected the card/s reach 99%. Sometimes even having better frames.
As for my system specs:
CPU: 4930k
GPU: Title.
HDD: 500GB Samsung SSD
RAM: 16GB 2133Mhzw Corsair Dominator
MOB: RIVE IV
Do you guys think that I am actually starting to hit a cpu bottleneck with my 4930k ? I am seeing my usage that low in most games I have tested: R6S, BF1 those are the games that have the nicest fidelity.
I am thinking about upgrading my cpu/mob and ram. What do you guys think ?
ALSO: Wait for the 9900k or go for the ryzen 2700 ?
Kind Regards
Toby
I think that's pretty good for an older gen you're using. Of course you're seeing seeing less than the peak performance (the loosely defined bottleneck) from the GPU's. The thing is, is it enough for you to break out the wallet? Personally and folks might take issue with this, unless you're a content creator that needs a bunch of slower multi-threaded cores, get the fastest 6-8 core intel part you can afford.
EVGA X99 FTWK / i7 6850K @ 4.5ghz / RTX 3080Ti FTW Ultra / 32GB Corsair LPX 3600mhz / Samsung 850Pro 256GB / Be Quiet BN516 Straight Power 12-1000w 80 Plus Platinum / Window 10 Pro
|
Tharifas88
New Member
- Total Posts : 5
- Reward points : 0
- Joined: 2011/03/28 12:20:07
- Status: offline
- Ribbons : 0
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/21 16:04:58
(permalink)
HeavyHemi
I think that's pretty good for an older gen you're using. Of course you're seeing seeing less than the peak performance (the loosely defined bottleneck) from the GPU's. The thing is, is it enough for you to break out the wallet? Personally and folks might take issue with this, unless you're a content creator that needs a bunch of slower multi-threaded cores, get the fastest 6-8 core intel part you can afford.
Well yeah I mean , it is a pretty old cpu I mean it was released in september 2013 IF I recall correctly. So its pushing 5 years. But I do think that the single core performance is showing cracks. Also the small things like ddr3 that I am using in comparison to ddr4 with higher clock speeds doesnt make much of a difference but it adds up. Lets not forget the fact that newer mobos have newer controller chips. I think all in all I need a new "base" for my gpu's.
post edited by Tharifas88 - 2018/09/21 16:13:20
|
arestavo
CLASSIFIED ULTRA Member
- Total Posts : 6916
- Reward points : 0
- Joined: 2008/02/06 06:58:57
- Location: Through the Scary Door
- Status: offline
- Ribbons : 76
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/21 23:43:25
(permalink)
With SLI, especially 1080 Tis, a higher resolution in SLI supported games means higher GPU utilization.* (depending on the game - some won't be helped even running at 8K because they don't take much processing power) Even in synthetic benchmarks like Firestrike (not Ultra, the 1080P version), which is designed to max out systems, I am lucky to get 75 to 90% GPU utilization. On the other hand, with the Ultra (4K) version of Firestrike and I hit 99% on both 1080 Tis. Things that will help include: using a HB SLI bridge (I've just seen this happen, coming from a flexible original SLI bridge to a HB bridge), and using Nvidia's DSR to artificially increase the resolution. You can also start playing DX12 games since they utilize more than one primary thread (DX11 really holds back performance with high powered GPUs and lower resolutions). Yes, getting a new system with a better IPC & higher OC CPU and faster RAM will help as well (but the next GPU upgrade you'll be right back where you were before).
post edited by arestavo - 2018/09/21 23:50:17
|
Tharifas88
New Member
- Total Posts : 5
- Reward points : 0
- Joined: 2011/03/28 12:20:07
- Status: offline
- Ribbons : 0
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/22 00:51:00
(permalink)
arestavo
With SLI, especially 1080 Tis, a higher resolution in SLI supported games means higher GPU utilization.* (depending on the game - some won't be helped even running at 8K because they don't take much processing power)
Even in synthetic benchmarks like Firestrike (not Ultra, the 1080P version), which is designed to max out systems, I am lucky to get 75 to 90% GPU utilization. On the other hand, with the Ultra (4K) version of Firestrike and I hit 99% on both 1080 Tis.
Things that will help include: using a HB SLI bridge (I've just seen this happen, coming from a flexible original SLI bridge to a HB bridge), and using Nvidia's DSR to artificially increase the resolution. You can also start playing DX12 games since they utilize more than one primary thread (DX11 really holds back performance with high powered GPUs and lower resolutions). Yes, getting a new system with a better IPC & higher OC CPU and faster RAM will help as well (but the next GPU upgrade you'll be right back where you were before).
As I stated in a previous post. I am well aware that lower resolutions dont max out thte gpu. BUT I mean come on BF1 at 1440 is absolutly taxing, as is siege. Its not like I am playing roblox with the damn cards ( I also like to play those games at higher frames , rather then at higher resolutions). Thats why I think upgrading my core components makes more sense. They are 5+ years old and I am sure that will make a big difference maybe 15-25% in performance I dont even think that , that is wishfull thinking either. I will also look into the HB Bridges had my eye on one for a while never bought it though becaus the reviews "I" read always seemed to vary, and or didnt have that much of a performance hike. Out of your experience using it what was your "net" gain after coming from a flex ? I am also using a single flex sli bridge atm. Or should I sell my two cards and just straight up go for 2080ti ? Haven heard that many good reviews about it, especially the money to performance ratio seems to be really bad. Thanks mate.
|
arestavo
CLASSIFIED ULTRA Member
- Total Posts : 6916
- Reward points : 0
- Joined: 2008/02/06 06:58:57
- Location: Through the Scary Door
- Status: offline
- Ribbons : 76
Re: 1080ti SLI low usage 70%"ish" on average
2018/09/22 08:44:11
(permalink)
Tharifas88
arestavo
With SLI, especially 1080 Tis, a higher resolution in SLI supported games means higher GPU utilization.* (depending on the game - some won't be helped even running at 8K because they don't take much processing power)
Even in synthetic benchmarks like Firestrike (not Ultra, the 1080P version), which is designed to max out systems, I am lucky to get 75 to 90% GPU utilization. On the other hand, with the Ultra (4K) version of Firestrike and I hit 99% on both 1080 Tis.
Things that will help include: using a HB SLI bridge (I've just seen this happen, coming from a flexible original SLI bridge to a HB bridge), and using Nvidia's DSR to artificially increase the resolution. You can also start playing DX12 games since they utilize more than one primary thread (DX11 really holds back performance with high powered GPUs and lower resolutions). Yes, getting a new system with a better IPC & higher OC CPU and faster RAM will help as well (but the next GPU upgrade you'll be right back where you were before).
As I stated in a previous post. I am well aware that lower resolutions dont max out thte gpu. BUT I mean come on BF1 at 1440 is absolutly taxing, as is siege. Its not like I am playing roblox with the damn cards ( I also like to play those games at higher frames , rather then at higher resolutions).
Thats why I think upgrading my core components makes more sense. They are 5+ years old and I am sure that will make a big difference maybe 15-25% in performance I dont even think that , that is wishfull thinking either.
I will also look into the HB Bridges had my eye on one for a while never bought it though becaus the reviews "I" read always seemed to vary, and or didnt have that much of a performance hike. Out of your experience using it what was your "net" gain after coming from a flex ? I am also using a single flex sli bridge atm.
Or should I sell my two cards and just straight up go for 2080ti ? Haven heard that many good reviews about it, especially the money to performance ratio seems to be really bad.
Thanks mate.
It depends on the games - but I saw The Witcher 3 (at 4K Ultra) go from 57-60 FPS to 100-103 FPS in a Nilfgaard camp. That's the exceptional case, NOT the rule. Even a benchmark like Timespy saw gains, and as a synthetic benchmark that was maxing out my GPU usage with a SLI ribbon cable that's impressive. Obviously, if the games that you play don't support SLI then the 2080 Ti will be better. BUT! At 1440P the 2080 Ti showed less gains over a 1080 Ti (because the rest of the system can't push the 2080 Ti to it's max all the time). And yes, you are paying out the nose for those gains. The value is yours to call.
|