| Page 233 of 782 |
|
|
Posted: Thu, 18th Dec 2014 22:30 Post subject: |
|
 |
Wonder if my 600w psu would handle it tho
5080 | ps5 pro
Sin317-"im 31 years old and still surprised at how much shit comes out of my ass actually ..."
SteamDRM-"Call of Duty is the symbol of the true perfection in every aspect. Call of Duty games are like Mozart's/Beethoven's symphonies"
deadpoetic-"are you new to the cyberspace?"
|
|
| Back to top |
|
 |
|
|
Posted: Thu, 18th Dec 2014 22:36 Post subject: |
|
 |
| KillerCrocker wrote: | | Wonder if my 600w psu would handle it tho |
-.- ofcourse, unless its some china bomb psu which gives 50 real watts. On a quality 600W PSU you can easily run SLI setup with overclocks along with overclocked i7 + other stuff.
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
|
|
Posted: Thu, 18th Dec 2014 23:14 Post subject: |
|
 |
I have tested out overclocked GTX 780 SLI on this Superflower 550W platinum without any problems, but this is tho absolutely top of the line PSU, can deliver like 650W on lower rating (Gold).
|
|
| Back to top |
|
 |
|
|
Posted: Fri, 19th Dec 2014 15:48 Post subject: |
|
 |
All these options now in nvidia control panel and what not its been a while but at least i can mess with the x034aadad34z000 < whatever options > for AA on games that hate u forcing it.
card is awesome, so well made.
p.s shader cache on disk wtf?
|
|
| Back to top |
|
 |
|
|
Posted: Sat, 20th Dec 2014 23:59 Post subject: |
|
 |
Hi All,
I'm considering getting a single 970 with the aim to SLI in the next few months.
I currently have a 670 SLI (2GB) setup and wonder what people think the impact would be?
From what I gather, the 970 performs pretty much on par with a my current setup but I get the added benefit of 4GB and games not optimised for SLI.
I don't have the funds to SLI two new cards now but I figured getting the single card now shouldn't really have any negative impact?
What do you think, waste of money? I think 4GB VRAM is going to be a big impact in the next 6-12 months? I'm hoping to have the SLI in time with Skylake.
Also, Just a side question.
Does anyone use the DSR setting got Nvidia cards, do you see any difference in image quality?
|
|
| Back to top |
|
 |
|
|
Posted: Sun, 21st Dec 2014 17:56 Post subject: |
|
 |
when zoomed in through a scope in game at long ranges, gpu will max the usage almost, this seems weird but common ive found with other cards I have owned, but what is this power limit spike doing when I get to max gpu usage end of the scale.
It has a range of 0 - 1 power limit in monitor it seems and only goes to 1 when at high gpu usage.
Also it seems backwards in dayz lol, when u are at wide fov and lots on screen the gpu usage is lower than when only in scope and further away rendering smaller view area
but the performance does not faulter when back to normal view, card is very good at keeping smooth play but some of the sense of it just baffles me.
here is a monitor shot of the power limit spike to 1 from 0 when the gpu usage was at 96 etc.
http://prntscr.com/5jco14 monitor pic
http://prntscr.com/5jcq70 view pic at reduced usage, trees in far hill infront of me with scope full zoom puts it at 96% usage
edit: hmm thinking about it I guess it means the card is reaching ones set power setting of 100% in the gpu tools, and 1 indicates its functioning as set. as i see it go to 1 when power output is at the high to near 100 mark.
i guess if i had the power setting to 130 or more and the card was not overclocked that power limit would not show 1 and stay at 0 as it would not be reaching my settings.
meh think thats it. I can't find any info on the monitor and its relations and tool tips are useless.
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
JBeckman
VIP Member
Posts: 35867
Location: Sweden
|
Posted: Mon, 22nd Dec 2014 12:46 Post subject: |
|
 |
http://www.guru3d.com/files-details/msi-afterburner-beta-download.html
MSI Afterburner 4.1.0 is out.
| Quote: |
Revision history - Version 4.1.0
Brand new user interface skins by Drerex Design. New skins use improved skin format introduced in the previous version of MSI
Afterburner skin compiler, which reduces each compiled skin size by factor of 10 on average. Legacy MSI Afterburner v2 and v3 skins
were also recompiled in new format and included in distributive for those who prefer old-style user interface
Improved hardware monitoring module, new robust hardware monitoring features have been ported from original RivaTuner
guru3d hardware monitoring module:
Added integrated viewer for native RivaTuner Hardware Monitoring Log (*.HML) files. Now you can view hardware
monitoring history stored in HML files either by pressing “View” button in “Monitoring” tab or directly open HML files via
Windows explorer
Added custom tracking markers support. Now you can hold <Alt> key while clicking hardware monitor window to set up
to 8 custom tracking markers in desired positions
Now you can customize graph color in layered mode via double clicking color legend box on the graph
Now tracking markers display current values for all graphs in layered mode instead of active graph only
Improved graphs resizing implementation for attached hardware monitor window
Improved skin browser:
Now “User interface” tab open immediately without waiting for processing of all installed skin files, which can be time
consuming if you have a lot of skin files. The list of installed skins is now being populated in separate background thread
Added precaching for skin preview images. Skin files are no longer being accessed when switching between skins in the
browser
Improved skin rendering performance for layered skin composition mode with alpha channel
Improved rendering performance for skins using animated indicators (e.g. animated profile save indicator in default MSI
Afterburner skin)
Added backbuffering support to skinned windows to simplify implementation of various framebuffer processing related
effects (e.g. skin scaling)
Vector indicators are now being rendered via GDI+ to provide improved antialiasing and rendering performance
Added support for obfuscated USF format, which is represented by some unfair competitors as 100% in-house
development. Now you can use MSI Afterburner’s built-in skin decompiler to analyze source code of such skin files
Added skins scaling support. The scaling may improves the appearance of some small third party skins on high resolution
monitors. Please take a note that upscaling may reduce the performance a bit
Improved skin cross-compatibility layer:
Extended list of applications supported by skin cross-compatibility layer. Now MSI Afterburner supports USF skins created
for any version of third party overclocking tools starting from minimalist oldschool skins from GeForce 200 series
overclocking era and ending by obfuscated USF skins used in modern overclocking tools
Added support for 10 profile slots in skin cross-compatibility mode
Now it is possible to open advanced settings window with <Ctrl> + <S> keyboard shortcut to provide compatibility with
third party skins having no dedicated button for accessing advanced settings window
Now it is possible to open info window with <Ctrl> + <I> keyboard shortcut to provide compatibility with third party skins
having no dedicated button for accessing info window
Added configuration file switch allowing power users to invert colors in text mode for Logitech keyboard LCD displays
Added silent installation scenario support
Ukrainian localization has been seriously revamped
RivaTuner Statistics Server has been upgraded to v6.3.0
|
(And where the changes for RivaTuner Statistics Server 6.3.0 should be it instead repeats the changes for MSI Afterburner.)
|
|
| Back to top |
|
 |
|
|
Posted: Mon, 22nd Dec 2014 12:56 Post subject: |
|
 |
@JBeckman
They say to uninstall AB before installing this new one, does that mean wiping profiles too?
|
|
| Back to top |
|
 |
|
|
Posted: Tue, 23rd Dec 2014 11:14 Post subject: |
|
 |
|
|
|
| Back to top |
|
 |
JBeckman
VIP Member
Posts: 35867
Location: Sweden
|
Posted: Fri, 26th Dec 2014 13:56 Post subject: |
|
 |
http://forums.guru3d.com/showthread.php?t=395714
| Quote: |
20nm and 16nm GPUs by AMD and Nvidia respectively have reportedly been delayed. Not due to any technical reason other than the limited capacity of both nodes. I’ve actually waited on this information for a few days to affirm its legitimacy. My sources have confirmed (again) that AMD’s upcoming graphics architecture is indeed 20nm based and that’s precisely the reason why it’s been slightly delayed.
Although I can’t link to the source as it’s not public, suffice to say this is the same source that was the first to tell us about Nvidia deciding to make GM204 ( GTX 980 and 970 ) using the 28nm instead of 20nm. This is also the same source that told us that Nvidia had skipped 20nm and went straight to 16nm designs.
No 16nm GPUs From Nvidia or AMD Next Year – 20nm GPUs Also Slightly Delayed
The reason for this delay is due to mobile chip makers, namely Apple and Qualcomm, gobbling up the limited early 20nm capacity. It’s not good news for 16nm either as the limited capacity of that node, especially in its early months of life, has also reportedly been drained almost dry by the same mobile giants.
Just like Qualcomm and Apple we may see Nvidia introduce 16nm Tegra mobile SOCs in late 2015. This development is interesting in the sense that it may drive Nvidia to cancel the 16nm Maxwell shrinks and go straight to Pascal instead for 2016. Either way, unsurprisingly there won’t be any 16nm GPUs until 2016. Even early 2016 is unlikely at this point based on TSMC’s own projected 16nm wafer revenue. However we don’t expect Nvidia to remain idle. The company still has one last ace to play in 2015 and that’s GM200.
If you remember back in September we had reported that the first of AMD’s 20nm R9 300 series graphics cards would arrive between mid February and early March of 2015. The delay has shifted the schedule by about two months. So the same 20nm chips would now arrive sometime between April and May of next year.
This however does not mean that AMD would not introduce any new products before April next year. As we’ve reported earlier, we’re likely to see both 20nm and 28nm graphics products from AMD in 2015. The latter will serve the market before the introduction of 20nm silicon. We should see the enhanced Hawaii version introduced along with a fully unlocked Tonga GPU before the flagship based on Fiji XT takes the stage.
The upcoming graphics architecture is code named Caribbean Islands. The amusing coincidence happens to be that the Caribbeans have infamously been known to be pirate islands (Jack Sparrow springs to mind). This play on words is interesting. And I suspect it’s been established to distinguish the Pirate Islands architecture (Bermuda, Fiji and Tonga) from a more distinct group of Caribbean Islands based GPUs. AMD likely made some significant architectural advancements that warranted a change in the code name.
All will become clear once AMD begins divulging the details of the graphics refresh it has planned for 2015.
|
Nothing confirmed yet but it will be interesting to see what AMD and Nvidia have planned for 2015
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
|
|
Posted: Fri, 26th Dec 2014 16:28 Post subject: |
|
 |
Amd is really slacky recently. 2 years ago their 78xx and 79xx cards were a bang for the buck but now ? I would get gtx970
5080 | ps5 pro
Sin317-"im 31 years old and still surprised at how much shit comes out of my ass actually ..."
SteamDRM-"Call of Duty is the symbol of the true perfection in every aspect. Call of Duty games are like Mozart's/Beethoven's symphonies"
deadpoetic-"are you new to the cyberspace?"
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
|
|
Posted: Fri, 26th Dec 2014 17:20 Post subject: |
|
 |
Why not? Its fast and efficient, dont see anything groundbreaking coming in next year, unless you want to pay half more (Titan 2 or something like that shit).
|
|
| Back to top |
|
 |
|
|
Posted: Fri, 26th Dec 2014 22:13 Post subject: |
|
 |
Assuming that the architecture stays similar, what effect could the new node have on performance?
|
|
| Back to top |
|
 |
Werelds
Special Little Man
Posts: 15098
Location: 0100111001001100
|
Posted: Fri, 26th Dec 2014 23:31 Post subject: |
|
 |
| Quote: | | The reason for this delay is due to mobile chip makers, namely Apple and Qualcomm, gobbling up the limited early 20nm capacity |
This should set off alarm bells with anyone who knows anything about these nodes. These things don't share the same production lines.
On 28nm, there are 5 lines (at TSMC): High Performance (HP), HPM (High Performance Mobile), HPL (High Performance Low [Power]) and LP (Low Power) and most recently, High Performance Compact (HPC). HP is what GPUs use. SoCs (Apple and Qualcomm) use HPC and HPM for low-end and high-end respectively - reason being that the HP line leaks too much to be used for SoCs. HPL and LP are meant for other applications, like cell radio chips, WiFi chips and the likes (Broadcom gobble up most of that that capacity).
At 20nm TSMC only has a SoC line capable of high volumes right now - there is no equivalent of the HP line that is capable of high volumes. Someone claiming that Apple/Qualcomm have all the 20nm capacity is correct, but that is completely irrelevant for GPUs. That's also why there's been rumours floating around about AMD using GF to produce lower end GPUs at 20nm; they wouldn't need crazy volumes for those so GF could offer enough capacity. For high-end they'd need more capacity than GF has to offer at present (although with GF's recent acquisitions and expansions, they're catching up to TSMC pretty quickly). And that's also the same reason why there's been rumours of both Nvidia and AMD skipping 20nm in favour of 16FF for their high-end GPUs.
Technically speaking the above isn't 100% accurate, as there aren't really separate "lines" (so don't picture it as separate conveyor belts or something along those lines ) but this is the easiest way to explain it. It's not caused by Apple or Qualcomm gobbling up capacity, it's caused by TSMC being incapable of producing HP HKMG chips in high volumes. Or rather, producing them with enough -if any- benefits over 28nm HP HKMG, which means it would be cost-ineffective for them to do it right now.
As for what new nodes do: more compact chips, lower power. So more transistors at higher frequencies. To give you an idea: we've gone from 3.73m transistors per mm^2 on 55nm at 650-750 MHz, to 6.79m/mm^2 on 40nm at 800-900 MHz, to 14.16m/mm^2 on 28nm anywhere from 1 GHz all the way to 1.4 GHz. Transistor density are all the highest achieved on each node, clock speeds are the minimum/maximums on either side. And in case anyone is wondering: it's all AMD for the transistor density (RV770, Cayman and Hawaii respectively, Nvidia generally sits about 1 million lower per mm^2, 13.06m/mm^2 for Maxwell is their best yet, mostly because AMD has more compact crossbars and memory controllers). Clock speeds differ, as not all architectures are equal and not all require the same speeds to achieve similar results, which is why I took min/max from both sides.
|
|
| Back to top |
|
 |
Morphineus
VIP Member
Posts: 24939
Location: Sweden
|
Posted: Mon, 29th Dec 2014 11:23 Post subject: |
|
 |
So I'm close to ordering a 970 and wondering which variant I should go for (reading up on it, but would like to hear the opinion of you guys).
I was leaning towards the Asus Strix, with my eye on EVGA (tho that one had iffy heatpipes if I remember correctly) as well. The gigabyte should fit in my case but I'm more worried about the flimsiness (build quality) as the weight of it's coolers actually have an effect on it (and while I always liked the brand it has a certain cheapness compared to others).
I'm not really looking for the highest 'clock' out there, since the difference is usually negligible. Built quality is high on my list, I won't be overclocking it in the near future. Will run how it comes out of the box with clocking later down the line if I have a need for that.
So which would you recommend. Price seems to be about the same around here, with only the Pallit having a real difference.
Edit: and it seems they are selling the MSI one after all (didn't mention it since I didn't see it in the store)

|
|
| Back to top |
|
 |
|
|
Posted: Mon, 29th Dec 2014 12:50 Post subject: |
|
 |
i'd go with strix, only 1x8 pin and a backplate, msi is 1x6 pin and 1x8 pin and no backplate, evga is nothing special and overall looks cheap, palit has thinner pcb and overall cooler quality is kind of crapy and wobly
|
|
| Back to top |
|
 |
Morphineus
VIP Member
Posts: 24939
Location: Sweden
|
Posted: Mon, 29th Dec 2014 12:58 Post subject: |
|
 |
| Mortibus wrote: | | i'd go with strix, only 1x8 pin and a backplate, msi is 1x6 pin and 1x8 pin and no backplate, evga is nothing special and overall looks cheap, palit has thinner pcb and overall cooler quality is kind of crapy and wobly |
Aye, so far the 8 pin and backplate is winning me over when I read up on everything. Just seems to me it's the better build of them all.
|
|
| Back to top |
|
 |
|
|
Posted: Mon, 29th Dec 2014 13:49 Post subject: |
|
 |
Well I can tell you as I am informed of late by plenty of reading, go with any of them if you do not intend to overclock like a banshee...
The only bonuses are marginal between cards at default clocks and boost and then you have the back plate and power choices to make after that only.
If though you intend to overclock a fair bit later on then go with a Gigabyte/msi/strix in that order as it stands now with bios and those that are making altered bios updates for the cards.
I have a strix its all good but nvidia drivers right now are lacking in certain games and older games that are in dx9 seem to be crashing more often. The latest drivers went from beta to whql in like days and although I got amazing performance from them they were just not stable in games I was playing.
Im sure when things settle down more with more driver releases and possible bios updates it wont matter which card you own in the 970 range, they all seem to do well.
|
|
| Back to top |
|
 |
Morphineus
VIP Member
Posts: 24939
Location: Sweden
|
Posted: Mon, 29th Dec 2014 14:23 Post subject: |
|
 |
It became a Strix in the end, not too bothered with having the best for clocking so it came down to: 1 wire less in my case and a sturdy backplate to seal the deal.
Really hate for how low my arcade cabinet went for... but at least I got meself a new GPU in the end.
Edit: Now it falls in the realm of waiting patiently until it arrives! (yeah right, patiently )
|
|
| Back to top |
|
 |
|
|
Posted: Mon, 29th Dec 2014 16:44 Post subject: |
|
 |
Heh its weird card range these but they do work although dont understand some of its operation choices im not used to seeing 50% gpu use in a game fully maxed with 8xmsaa and everything else on still not taxing it. A fair bit is drivers I dont know how u force like the ati cards on nvidias to always run max 100% use, The adaptive and max power profile drop down seems to do nothing lol.
|
|
| Back to top |
|
 |
|
|
Posted: Tue, 30th Dec 2014 10:52 Post subject: |
|
 |
|
|
|
| Back to top |
|
 |
|
|
Posted: Tue, 30th Dec 2014 13:21 Post subject: |
|
 |
Overclock it to 980?
The stock clock is 1115 lol
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
|
|
Posted: Tue, 30th Dec 2014 15:09 Post subject: |
|
 |
Ah good someone that might be able to fill in this weird info for me then.
Ok I had noticed yesterday that my clock was without doing anything and set with no OC etc just defaults and it was running at 1114 on the monitor tools here.
Now I played all day fine with that and the card was boosting up to 1445mhz with power limit on 110 - I had set nothing else and I only set PL to 110 because I had noticed the PL monitor hitting 1 while playing and thought hmm maybe its trying or wanting more POWA! loool, so i gave 110 and nothing phased it was all good.
I log into pc today and my running clock while doing nothing is now saying 899 :O
which it was saying that other day before I uninstalled MSI 4.0.0 and installed 4.1.0
Tried resetting defaults in inspector/nvidia control panel/and on the MSI Afterburner. I use afterburner and not GPU Tweak for my asus card because I dont like how the monitor tools limits view size and you have to scroll the window to see the graphs.
so its sitting at 899 core clock and 3505mhz memory with the asus gtx 970 strix dcII right now which I think is probably normal for 3 monitor setup. I can drop it to 135 mhz I believe if using inspector to force 2d low clock settings.
all seems normal? and if you know why that original clock was sitting at 1114 as standard idle I would like to know
p.s I dont understand when you say you have clocked yours to 1253mhz when mine goes to that without even using any clock options. Here is pic of the boost that is natural to the card. the little red triangle top left on dial as you can see is in the 1250 area. http://prntscr.com/5mlvwa
edit: I see sorry "aka 980 boost base" is that really 1253 though when boosted? seems low, more I look the more confoosing your info seems, if your boost was set to 980 on the clock reading your gpu would not be making 1253 im sure, is the wording incorrect here of what you are saying?
This would be what setting the boost clock to 980 would look like, as you see the core running would be at little over 800 as you can see. To get a 980 boost clock setting you would have to declock the core by at least 275mhs
http://prntscr.com/5mlzyp
|
|
| Back to top |
|
 |
|
|
|
| Back to top |
|
 |
| Page 233 of 782 |
All times are GMT + 1 Hour |