Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [TT] NVIDIA should launch its next-gen Pascal GPUs with HBM2 in 2H 2016
New Posts  All Forums:Forum Nav:

[TT] NVIDIA should launch its next-gen Pascal GPUs with HBM2 in 2H 2016 - Page 32

post #311 of 724
Quote:
Originally Posted by Themisseble View Post

For sure.
hardwarecanucks just want to make GTX 780Ti look good. And thats ALL. There is now way that GTX 780Ti only 5-10% slower than GTX 980. While R9 390X is basically as fast as GTX 980.

In other hand R9 290X = R9 390.
I just hope they didnt use stock R9 290X in that test.

And here is Duderandom84 performance test
https://www.youtube.com/watch?v=-scxy-7qtMY
https://www.youtube.com/watch?v=YEtJHG27P0E
https://www.youtube.com/watch?v=bhXGyd3vs9E
https://www.youtube.com/watch?v=MAc7phqSdms
https://www.youtube.com/watch?v=y4D8qgDpFNc
- Dont tell that tessellation is the problem.. R9 280X is doing just fine against R9 380X.
- GTX 780Ti with 1150MHz core clock.

https://www.techpowerup.com/gpudb/2512/geforce-gtx-780-ti.html
So at least GTX 780Ti with actually great overclock isnt matching GTX 980... and hardwarecanucks says that GTX 780Ti is only 10% behind GTX 980.

Dude, I was there when the 980 launched. My OG Titan at 1300+MHz easily beat most 980's in the benching section for a long while after their release. Nowadays, of course, there's no comparison; my 1300MHz Titan can't even come close to the 970 today...
post #312 of 724
Re: AMD's launch driver being poorly done.

No. Plenty of us bought Hawaii because it was competitive in its cost bracket at launch. There were no broad issues with the drivers out of the gate. It's true that DX11 has been a steep learning curve for AMD to properly task it into the highly parallel GCN, which is where most of the gains over time come from, but that doesn't mean the launch drivers were gimped.

It competed at launch with the 780. It now competes with a completely different class of Nvidia GPU today due to scheduling optimizations. To spin that into a negative is ridiculous.
Parasite
(18 items)
 
  
CPUMotherboardGraphicsGraphics
i7 4770K @ 4.7GHz Z87 MPOWER (MS-7818) Sapphire Radeon 290x @1100/1500 EVGA 1080Ti SC2 Hybrid 
RAMHard DriveHard DriveCooling
G.SKILL 2133 Samsung 850 Pro Caviar Black Corsair H100 
CoolingCoolingOSMonitor
Corsair HG10 Corsair H60 Windows 7 x64 Sony XBR65X850B 
KeyboardPowerCaseMouse
CMSTORM Quickfire XT Corsair AX1200i Antec P280 Logitec G700 
Mouse PadAudio
Black, came with my NeXTcube 25 years ago. Sound Blaster Recon 3D PCIe 
  hide details  
Reply
Parasite
(18 items)
 
  
CPUMotherboardGraphicsGraphics
i7 4770K @ 4.7GHz Z87 MPOWER (MS-7818) Sapphire Radeon 290x @1100/1500 EVGA 1080Ti SC2 Hybrid 
RAMHard DriveHard DriveCooling
G.SKILL 2133 Samsung 850 Pro Caviar Black Corsair H100 
CoolingCoolingOSMonitor
Corsair HG10 Corsair H60 Windows 7 x64 Sony XBR65X850B 
KeyboardPowerCaseMouse
CMSTORM Quickfire XT Corsair AX1200i Antec P280 Logitec G700 
Mouse PadAudio
Black, came with my NeXTcube 25 years ago. Sound Blaster Recon 3D PCIe 
  hide details  
Reply
post #313 of 724
Quote:
Originally Posted by infranoia View Post

Re: AMD's launch driver being poorly done.

No. Plenty of us bought Hawaii because it was competitive in its cost bracket at launch. There were no broad issues with the drivers out of the gate. It's true that DX11 has been a steep learning curve for AMD to properly task it into the highly parallel GCN, which is where most of the gains over time come from, but that doesn't mean the launch drivers were gimped.

It competed at launch with the 780. It now competes with a completely different class of Nvidia GPU today due to scheduling optimizations. To spin that into a negative is ridiculous.

you're taking it the wrong way.

AMD isn't necessarily gimping their release drivers either, think of it as having blind spots on a new architecture, that they later on discover thus improving later drivers.
in this case, AMD found ways to drastically improve performance without changing the architecture itself, but also implies that they didn't find these things in the beginning, despite being designed by them.

on the other hand, i find Nvidia's a bit unusual, their later driver compared release driver have small improvements, they rarely make driver releases that continuously improve GPU speed.
this either means Nvidia's is slacking in driver improvements, or their initial driver release is just that good, or rather their driver team and architectural design team is more cooperative and coordinated.


forwarding these two points, i suspect Fury's issues regarding it's potential performance falling short of expectation is due to driver deficiencies, so we'd probably see it out-perform 980Ti later on.
but with Nvidia taking their turn, we might see what HBM can actually do much earlier than that, though that is if they've designed the architecture superbly, and their driver implementation being a good one.
Edited by epic1337 - 2/11/16 at 10:59pm
post #314 of 724
Quote:
Originally Posted by Majin SSJ Eric View Post

Dude, I was there when the 980 launched. My OG Titan at 1300+MHz easily beat most 980's in the benching section for a long while after their release. Nowadays, of course, there's no comparison; my 1300MHz Titan can't even come close to the 970 today...
but Titan's didn't even bench better than a 980 on release at stock... My SC 980 could easily outbench your Titan's on Release day?
 
4th time
(18 items)
 
3rd time
(11 items)
 
CPUMotherboardGraphicsRAM
5820k ASrock x99 Fatal1ty professional Gigabyte 1080 ti Corsair Vengeance DDR4 CAS 15 32GB 
Hard DriveCoolingOSMonitor
Samsung EVO 850 1TB SSD Corsair 110i GTX Microsoft Windows 10 Professional Acer 24 inch IPS 
MonitorMonitorMonitorKeyboard
Acer 24 inch IPS Acer 24 inch IPS Samsung 40 inch Corsair Strafe RGB MX cherry blue 
PowerCaseMouseMouse Pad
Corsair HX1000i Corsair Air 540 Corsair M65 RGB  Razer Manticor  
CPUMotherboardGraphicsRAM
i7 2600k ASROCK Fatal1ty p67 EVGA GTX 980 SC 16380 MBMBDDR3-SDRAM (Gskill) 1600mhz 
Hard DriveHard DriveCoolingOS
Samsung HD103SJ x 2 Samsung 850 EVO  Corsair H100i Windows 10 Professional 
MonitorMonitorMonitorMonitor
Samsung - LN40D550 - 40" LCD TV - 1080p ACER R240HY ACER R240HY ACER R240HY 
KeyboardPowerCaseMouse
razer black widow Corsair TX850 Coolermaster HAF 932 CM Storm Recon 
CPUMotherboardGraphicsRAM
955 MSI 870S-G54 2 x 6850 HIS IceQ-x 8 Gigabytes Gskill 1600 
Hard DriveCoolingOSMonitor
Samsung F3 1TB zalman 90mm Windows Ultimate x64 LG 47'  
MonitorPowerCase
Vizio 37' E370L  OCZ Modstream 700W Logisys CS888CL 
  hide details  
Reply
 
4th time
(18 items)
 
3rd time
(11 items)
 
CPUMotherboardGraphicsRAM
5820k ASrock x99 Fatal1ty professional Gigabyte 1080 ti Corsair Vengeance DDR4 CAS 15 32GB 
Hard DriveCoolingOSMonitor
Samsung EVO 850 1TB SSD Corsair 110i GTX Microsoft Windows 10 Professional Acer 24 inch IPS 
MonitorMonitorMonitorKeyboard
Acer 24 inch IPS Acer 24 inch IPS Samsung 40 inch Corsair Strafe RGB MX cherry blue 
PowerCaseMouseMouse Pad
Corsair HX1000i Corsair Air 540 Corsair M65 RGB  Razer Manticor  
CPUMotherboardGraphicsRAM
i7 2600k ASROCK Fatal1ty p67 EVGA GTX 980 SC 16380 MBMBDDR3-SDRAM (Gskill) 1600mhz 
Hard DriveHard DriveCoolingOS
Samsung HD103SJ x 2 Samsung 850 EVO  Corsair H100i Windows 10 Professional 
MonitorMonitorMonitorMonitor
Samsung - LN40D550 - 40" LCD TV - 1080p ACER R240HY ACER R240HY ACER R240HY 
KeyboardPowerCaseMouse
razer black widow Corsair TX850 Coolermaster HAF 932 CM Storm Recon 
CPUMotherboardGraphicsRAM
955 MSI 870S-G54 2 x 6850 HIS IceQ-x 8 Gigabytes Gskill 1600 
Hard DriveCoolingOSMonitor
Samsung F3 1TB zalman 90mm Windows Ultimate x64 LG 47'  
MonitorPowerCase
Vizio 37' E370L  OCZ Modstream 700W Logisys CS888CL 
  hide details  
Reply
post #315 of 724
Quote:
Originally Posted by infranoia View Post

Re: AMD's launch driver being poorly done.

No. Plenty of us bought Hawaii because it was competitive in its cost bracket at launch. There were no broad issues with the drivers out of the gate. It's true that DX11 has been a steep learning curve for AMD to properly task it into the highly parallel GCN, which is where most of the gains over time come from, but that doesn't mean the launch drivers were gimped.

It competed at launch with the 780. It now competes with a completely different class of Nvidia GPU today due to scheduling optimizations. To spin that into a negative is ridiculous.

As epic said, you're definitely taking it the wrong way. We're (or me at least) most definitely not saying AMD gimped their launch drivers, but rather since GCN's debut with 7970, they seem to somehow find some way to squeeze an extra 20% performance (give or take) out of their GPUs a few years down the road. Now whether this is because they're finally getting good at optimizing for GCN, or the launch driver team was asleep at the helm I don't know. BUT what I am saying is that if AMD launched their cards with that +20% already baked into the drivers, the competitive landscape in the last few years would've been very different, and perhaps AMD would still have around 40% market share.
Quote:
Originally Posted by epic1337 View Post

you're taking it the wrong way.

AMD isn't necessarily gimping their release drivers either, think of it as having blind spots on a new architecture, that they later on discover thus improving later drivers.
in this case, AMD found ways to drastically improve performance without changing the architecture itself, but also implies that they didn't find these things in the beginning, despite being designed by them.

on the other hand, i find Nvidia's a bit unusual, their later driver compared release driver have small improvements, they rarely make driver releases that continuously improve GPU speed.
this either means Nvidia's is slacking in driver improvements, or their initial driver release is just that good, or rather their driver team and architectural design team is more cooperative and coordinated.


forwarding these two points, i suspect Fury's issues regarding it's potential performance falling short of expectation is due to driver deficiencies, so we'd probably see it out-perform 980Ti later on.
but with Nvidia taking their turn, we might see what HBM can actually do much earlier than that, though that is if they've designed the architecture superbly, and their driver implementation being a good one.

Or maybe drivers already came maxed out from the factory. biggrin.gif
post #316 of 724
Quote:
Originally Posted by magnek View Post

Or maybe drivers already came maxed out from the factory. biggrin.gif
yes thats also quite likely, we know that Nvidia takes it's time before releasing their GPUs, so they have plenty of time improving drivers before release.
post #317 of 724
Quote:
Originally Posted by Darkpriest667 View Post

but Titan's didn't even bench better than a 980 on release at stock... My SC 980 could easily outbench your Titan's on Release day?

HERE: (posted day after NDA lifted and reviews posted)
Quote:
Originally Posted by OccamRazor View Post

Well, the new Maxwell core are out and are very good OC'ers but as everyone should know it's a bit of a side grade for

780Ti owners that have good cards with high OC's, Titan and 780 owners that also have good cards volt mod enabled because

wont be that much slower either!
And frankly only lower tier card owners (770 and below and lower than 290 in AMD side), bencher's and people that like to

stay in the edge will buy these cards, but the vast majority that have 780Ti/Titan Black/Titan/780 wont upgrade because

they will loose $$$ in the process and get a good card but only up to AVERAGE (best case scenario) 20% more performance,

reflected in FPS its close to 16 fps AVERAGE (best case scenario) from GK110!


(AVERAGE PERFORMANCE) GTX980@1450mhz = TitanBlack@1250mhz = GTX780Ti@1300mhz = Titan@1350mhz = GTX 780@1450mhz
Warning: Spoiler! (Click to show)


No soft voltmod for the GTX 970/980! The NCP81174 does not have support for I2C, all voltage control through driver up to

1,25V in Maxwell GPUs only!
Warning: Spoiler! (Click to show)










Red: Missing power phases (2)
Purple: 4 power phases for the core
Green: 1 power phase for the memory
Blue: Voltage controller NCP81174
Yellow: Missing capacitors and resistances front and back of card

Some of you might remember the GTX680 with a similar setting: Missing parts!
While this does not equate in bad performance in any way or puts the card in any harm, but it just leaves a bad taste in

your mouth after paying $$$ for your card...
If you intent to hard mod the card later on you probably will have to solder most of the missing caps! wink.gif


Its a wonderful card if you ask me: very fast, under water will be very cool, the Classified versions will be AMAZING!!!

it just leaves me asking a question: Where is GM210? tongue.gif
Just my 2 cents! thumb.gif


Cheers

Occamrazor

now mind you, this is before AIB partner cards hit the market and much longer before anyone started modding BIOSs getting core speed up to 1600 quite easily.
loon 3.2
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7-3770K Asus P8Z77-V Pro EVGA 980TI SC+ 16Gb PNY ddr3 1866 
Hard DriveHard DriveHard DriveOptical Drive
PNY 1311 240Gb 1 TB Seagate 3 TB WD Blue DVD DVDRW+/- 
CoolingCoolingOSMonitor
EKWB P280 kit EK-VGA supremacy Win X LG 24MC57HQ-P 
KeyboardPowerCaseMouse
Ducky Zero [blues] EVGA SuperNova 750 G2 Stryker M [hammered and drilled] corsair M65 
AudioAudio
SB Recon3D Klipsch ProMedia 2.1  
  hide details  
Reply
loon 3.2
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7-3770K Asus P8Z77-V Pro EVGA 980TI SC+ 16Gb PNY ddr3 1866 
Hard DriveHard DriveHard DriveOptical Drive
PNY 1311 240Gb 1 TB Seagate 3 TB WD Blue DVD DVDRW+/- 
CoolingCoolingOSMonitor
EKWB P280 kit EK-VGA supremacy Win X LG 24MC57HQ-P 
KeyboardPowerCaseMouse
Ducky Zero [blues] EVGA SuperNova 750 G2 Stryker M [hammered and drilled] corsair M65 
AudioAudio
SB Recon3D Klipsch ProMedia 2.1  
  hide details  
Reply
post #318 of 724
X99
(15 items)
 
  
CPUMotherboardGraphicsRAM
6950X Asrock X99 Taichi Lightning Z Corsair DP- 16gb-3000 
CoolingOSMonitorPower
960 Pro-500gb Windows 10  Acer Predator Z35 Evga 850 P2 
CaseMouseAudio
Corsair Air 540 G900 Sound blaster Z 
  hide details  
Reply
X99
(15 items)
 
  
CPUMotherboardGraphicsRAM
6950X Asrock X99 Taichi Lightning Z Corsair DP- 16gb-3000 
CoolingOSMonitorPower
960 Pro-500gb Windows 10  Acer Predator Z35 Evga 850 P2 
CaseMouseAudio
Corsair Air 540 G900 Sound blaster Z 
  hide details  
Reply
post #319 of 724
Thread Starter 
Quote:
Originally Posted by Maintenance Bot View Post

More good news.

http://wccftech.com/nvidia-pascal-graphics-cards-spotte








I'd say that definitely looks like an April launch.
MyCleanPC
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 4770k ASUS MAXIMUS VI EVGA master blaster Corsair Vengence  
Hard DriveOptical DriveOSMonitor
OCZ SSD raid0 samsung Win 7 Samsung 
PowerCase
Enermax rev 1050 Stacker 832 
  hide details  
Reply
MyCleanPC
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 4770k ASUS MAXIMUS VI EVGA master blaster Corsair Vengence  
Hard DriveOptical DriveOSMonitor
OCZ SSD raid0 samsung Win 7 Samsung 
PowerCase
Enermax rev 1050 Stacker 832 
  hide details  
Reply
post #320 of 724
Quote:
Originally Posted by Maintenance Bot View Post

More good news.

http://wccftech.com/nvidia-pascal-graphics-cards-spotted/


Exactly what makes these Pascal cards or even graphic cards from Nvidia?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [TT] NVIDIA should launch its next-gen Pascal GPUs with HBM2 in 2H 2016