• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

970's having performance issues using 4GB Vram - Nvidia investigating

Status
Not open for further replies.

Haha, that forum makes this one look like Peace & Love Central by comparison :D

Anyway, why hasn't anyone asked nV what the bandwidth of that final 500MB should be under normal operation? They should be able to give a figure.

There seems to be a lot of uncertainty as to whether this homebrew diagnostic tool is actually accessing the smaller VRAM pool or system memory... Have also heard that some is permanently reserved for PhysX, etc...
 
So, all those who are having game issues, can you tell me what games you are seeing it in please. I am happy to test and see what is what on a Titan. At least we can determine if it is indeed using more than 4GB. I am sure Kaap is happy to help out as well.

@ Kaap, Being a pain but can you test on your 290X machine as well for SoM please.

It will take a little while as I have to download it first so I will post in an hour or so.:)
 
Well if it runs fine on a 980 with no adverse effects using less than a Titan what does that tell you ?

That it doesn't need 4.5GB and why would it need conveniently the exact amount of VRAM 980 has? What is more likely is that it needs less than 4GB to run the benchmark (probably less than 3.5GB) and will fill up the remainder with cache/left over texture data or whatever up to a max of around 4.5GB.

Due to way Nvidia have created the memory setup on the 970 this means 3.5GB used.

No I think the 980 is struggling as well but just about gets away with it.

If you compare how this runs on the Titans to how it runs on the 980s it becomes obvious. On the 980s it is ok but on the Titans it is smoother than the 980s or 290Xs for that matter.:)

I think people are way to quick to write the 970 off as it is a very good card and people are over reacting.
 
I still don't quite get it... if the extra 0.5GB makes NO difference to performance, despite the way it's accessed, why have they done it exactly, and why not make it the same as the 980? Surely the only reason they would do so is to hobble the 970 so the 980 stands head and shoulders above as their flagship all singing and dancing wonder card... rather than sharing the stage with the 970, which it would surely do were it not for this 'design/fault' (delete as applicable) that they've implemented? I still half think that they realised that if they didn't do something, the 970 would be near level pegging with the 980.
 
I still don't quite get it... if the extra 0.5GB makes NO difference to performance, despite the way it's accessed, why have they done it exactly, and why not make it the same as the 980? Surely the only reason they would do so is to hobble the 970 so the 980 stands head and shoulders above as their flagship all singing and dancing wonder card... rather than sharing the stage with the 970, which it would surely do were it not for this 'design/fault' (delete as applicable) that they've implemented? I still half think that they realised that if they didn't do something, the 970 would be near level pegging with the 980.

Because of the way they've had to cut SMMs
 
If you compare how this runs on the Titans to how it runs on the 980s it becomes obvious.

You didn't post your benchmark results for the single Titan for me to make a comparison :p only the VRAM usage which was using more than 4GB 'because it can' which is what I was expecting.

I'm not writing off the 970 btw just stating that I think it will use 3.5GB - 4GB and show it in Afterburner but only if it actually needs the extra which in my opinion the Mordor bench does not.
 
Also Nai's benchmark is bugged, people need to stop using it

Thing is there are 980 users running that tool and getting the same results, which indicates there is certainly a problem with either the tool or understanding how to use it.

Or that there is another problem affecting both cards and being incorrectly ascribed to the 970's split memory.
 
Last edited:
You didn't post your benchmark results for the single Titan for me to make a comparison :p only the VRAM usage which was using more than 4GB 'because it can' which is what I was expecting.

I'm not writing off the 970 btw just stating that I think it will use 3.5GB - 4GB and show it in Afterburner but only if it actually needs the extra which in my opinion the Mordor bench does not.

Get someone to run max settings @4K then on a 970.:D
 
I still don't quite get it... if the extra 0.5GB makes NO difference to performance, despite the way it's accessed, why have they done it exactly, and why not make it the same as the 980? Surely the only reason they would do so is to hobble the 970 so the 980 stands head and shoulders above as their flagship all singing and dancing wonder card... rather than sharing the stage with the 970, which it would surely do were it not for this 'design/fault' (delete as applicable) that they've implemented? I still half think that they realised that if they didn't do something, the 970 would be near level pegging with the 980.

The GTX 970 has 1664 Cuda cores
The GTX 980 has 2048 Cuda cores

That is where it has been hobbled and from what I can see, nothing more. The memory reading is done differently on the 970 over the 980 because of the differing SMs and fewer crossbar resources on the memory system. It is all explained in ManuelG's post but I guess you don't want to see that answer and want something else instead.
 
The bench doesn't, no. The game however will and does call on more. Allocates almost all of TITAN's frame buffer at 1440p in fact. People should stop using Mordor as an example as it proves less than nothing.
 
Hi Gregster, could you please test your titans in SoM in the area of Nurn please? Specifically in the first part where there's slightly more grass and a you can fight about 3-4 orcs, the second area I noticed it in is in the Orc village in the North I believe.
Both times I noticed when vram went over 3.5gb I noticed a lot of stuttering.

In Dragon Age Inquisition I noticed it in the Hinterlands specifically the woods where you fight both mages and templars.
Other games I have noticed this in are Ass Creed Unity but like someone else has said this isn't the best example hence me testing it out on the other games.

I'd be happy to be proven wrong as I feel a little gutted in getting £600 worth of cards and would much rather prefer that I'm doing something stupid that's causing these issues than have to return the cards as otherwise they are awesome and the software feels much better to use than AMD's.

System specs if they help are the following:

4790K at stock
16gb G.Skill 2400Mhz I believe
OcUK Reference 970's SLI
Superflower 1000w Platinum psu


Thanks
 
Wow so the 980 (not just the 970) lose performance as well when using above 3.5GB vram "by design" according to the Nvida mod? I better stop recommending people to consider getting the 980 instead of the 970...

Can you not understand that when you dial up the settings to use the extra VRAM that frames will drop? I thought that was basic common sense.
 
i think trying to recreate in a game would be 100 to 1 anyway if you only get the slowdown from heavy reading on the last 500mb, not just caching
but u kind of going around in circles here tbh :o
 
Its quite simple - the GTX970 does not have the fully enabled memory controller resources the GTX980 has it appears.

TR were the first site to notice something was bit off last year:

http://techreport.com/blog/27143/here-another-reason-the-geforce-gtx-970-is-slower-than-the-gtx-980

upEVZl1.gif


TBH,its not like the Nvidia second tier top cards have not had slightly less RAM or bandwidth than the top end ones before. I don't think it will make much bearing on whether you would buy a GTX970 or not TBH.
 
Hi Gregster, could you please test your titans in SoM in the area of Nurn please? Specifically in the first part where there's slightly more grass and a you can fight about 3-4 orcs, the second area I noticed it in is in the Orc village in the North I believe.
Both times I noticed when vram went over 3.5gb I noticed a lot of stuttering.

In Dragon Age Inquisition I noticed it in the Hinterlands specifically the woods where you fight both mages and templars.
Other games I have noticed this in are Ass Creed Unity but like someone else has said this isn't the best example hence me testing it out on the other games.

I'd be happy to be proven wrong as I feel a little gutted in getting £600 worth of cards and would much rather prefer that I'm doing something stupid that's causing these issues than have to return the cards as otherwise they are awesome and the software feels much better to use than AMD's.

System specs if they help are the following:

4790K at stock
16gb G.Skill 2400Mhz I believe
OcUK Reference 970's SLI
Superflower 1000w Platinum psu


Thanks

I will do and one thing I did notice when I played and completed SoM was even on 3 Titans that had plenty of VRAM, I would get an occasional drop in frames that resulted in a stutter in heavy areas. I played it to death and even after fully optimised drivers, I still got the stutter in certain places.

I feel that some of the latest games and especially UBI games are just plain bugged and not very good.

I will rerun it though :)
 
Status
Not open for further replies.
Back
Top Bottom