• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Next generation GPU jump how big really it will be on this time?

Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
At first let's look at some historical facts about performance gain over GPU generations.

All information is from TechPowerUp benchmark tests (performance summary comparison).

From GTX480 (40nm) to GTX580 (40nm) performance gain ~ 16%
From GTX580 (40nm) to GTX680 (28nm) performance gain ~ 30%
From GTX680 (28nm) to GTX780 (28nm) performance gain ~ 21%
From GTX780 (28nm) to GTX980 (28nm) performance gain ~ 30%
From GTX980 (28nm) to GTX1080 (16nm) performance gain ~ 67%
From GTX1080 (16nm) to GTX1180 (12nm) performance gain ~ 35 to 45% ??? (If look at current rumored specs) 3584 Cuda Cores with GDDR6 memory and probably the same Volta architecture but only without tensor cores.

I will remind that GTX1080TI is ~ 35% faster than GTX1080
And Titan V with 43% more Cuda Cores and HBM2 memory (Volta architecture) is only on average ~ 25% faster than GTX1080TI (That information is form PC Perspective benchmark test).

What do you think about that ?
 
Last edited:
Joined
Feb 8, 2012
Messages
3,012 (0.68/day)
Location
Zagreb, Croatia
System Name Windows 10 64-bit Core i7 6700
Processor Intel Core i7 6700
Motherboard Asus Z170M-PLUS
Cooling Corsair AIO
Memory 2 x 8 GB Kingston DDR4 2666
Video Card(s) Gigabyte NVIDIA GeForce GTX 1060 6GB
Storage Western Digital Caviar Blue 1 TB, Seagate Baracuda 1 TB
Display(s) Dell P2414H
Case Corsair Carbide Air 540
Audio Device(s) Realtek HD Audio
Power Supply Corsair TX v2 650W
Mouse Steelseries Sensei
Keyboard CM Storm Quickfire Pro, Cherry MX Reds
Software MS Windows 10 Pro 64-bit
You are looking at performance gains between generations completely out of context.
You have massive architectural change first fermi to kepler (no more double clocked shaders), then from kepler to maxwell (tile based rasterizer), then maxwell to pascal massive clock change ... that's proper context for those increases and a clue for next gen ;)
 
Joined
Sep 17, 2014
Messages
20,692 (5.96/day)
Location
The Washing Machine
Processor i7 8700k 4.6Ghz @ 1.24V
Motherboard AsRock Fatal1ty K6 Z370
Cooling beQuiet! Dark Rock Pro 3
Memory 16GB Corsair Vengeance LPX 3200/C16
Video Card(s) ASRock RX7900XT Phantom Gaming
Storage Samsung 850 EVO 1TB + Samsung 830 256GB + Crucial BX100 250GB + Toshiba 1TB HDD
Display(s) Gigabyte G34QWC (3440x1440)
Case Fractal Design Define R5
Audio Device(s) Harman Kardon AVR137 + 2.1
Power Supply EVGA Supernova G2 750W
Mouse XTRFY M42
Keyboard Lenovo Thinkpad Trackpoint II
Software W10 x64
Generally the jump is 25-30% per gen. That 67% over there is wildly exaggerated because we now put the 980 in resolutions and situations where it runs out of VRAM or shaders, such as 4K, while a 1080 does fine at that res. When you restrict the comparison to the mainstream resolution (1080p), it becomes a whole other reality. Agreed, for Maxwell > Pascal the jump is still over 30%, but rather in the region of 35-40% at best.
 
Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
And Titan V with 43% more Cuda Cores and HBM2 memory (Volta architecture) is only on average ~ 25% faster than GTX1080TI (That information is form PC Perspective benchmark test).

This information to me personally sounds very bad! Because that's is GPU refresh at the best! There is no IPC improvement if you compare cuda core count. Two different architectures really? More looks like shrink down form 16nm to 12nm and some tweaks.
 
Last edited:
Joined
May 8, 2016
Messages
1,730 (0.60/day)
System Name BOX
Processor Core i7 6950X @ 4,26GHz (1,28V)
Motherboard X99 SOC Champion (BIOS F23c + bifurcation mod)
Cooling Thermalright Venomous-X + 2x Delta 38mm PWM (Push-Pull)
Memory Patriot Viper Steel 4000MHz CL16 4x8GB (@3240MHz CL12.12.12.24 CR2T @ 1,48V)
Video Card(s) Titan V (~1650MHz @ 0.77V, HBM2 1GHz, Forced P2 state [OFF])
Storage WD SN850X 2TB + Samsung EVO 2TB (SATA) + Seagate Exos X20 20TB (4Kn mode)
Display(s) LG 27GP950-B
Case Fractal Design Meshify 2 XL
Audio Device(s) Motu M4 (audio interface) + ATH-A900Z + Behringer C-1
Power Supply Seasonic X-760 (760W)
Mouse Logitech RX-250
Keyboard HP KB-9970
Software Windows 10 Pro x64
For architecture, compare fully unlocked ones.
So GTX 580 to GTX 780 Ti to Titan XP (ie. full big chips to full big chips), or GTX 560 Ti to GTX 680 to GTX 980 to GTX 1080.
 
Last edited:
Joined
May 25, 2009
Messages
236 (0.04/day)
Processor AMD R5 7600X
Motherboard Asrock X670E Pro RS
Cooling Noctua NH-15S
Memory 2*16 GB 5600 CL34
Video Card(s) XFX 6800XT 319 Merc
Storage Samsung 970 Evo
Power Supply Super Flower 850 Gold
Mouse Steelseries 310
And Titan V with 43% more Cuda Cores and HBM2 memory (Volta architecture) is only on average ~ 25% faster than GTX 1080 TI (That information is form PC Perspective benchmark test).

This information to me personally sounds very bad! Because that's is GPU refresh at the best! There is no IPC improvement if you compare cuda core count. Two different architectures really?

Yeah but clock were low since the half node is hardly mature, the GPU is giant (800mm^2 is a lot even for a seasoned process), and the V brings the Titans back to their semi-professional roots, meaning a 2080/1180Ti would've done better in gaming, having a bigger TDP, more gaming oriented BIOS and driver, not to mention no tensor cores, and other Compute hw tweaks. But yeah, unless Turing brings some gaming enhancements, and not just cut the pro elements, its gonna be a second gen of 0 IPC
 
Joined
Jan 8, 2017
Messages
8,810 (3.35/day)
System Name Good enough
Processor AMD Ryzen R9 7900 - Alphacool Eisblock XPX Aurora Edge
Motherboard ASRock B650 Pro RS
Cooling 2x 360mm NexXxoS ST30 X-Flow, 1x 360mm NexXxoS ST30, 1x 240mm NexXxoS ST30
Memory 32GB - FURY Beast RGB 5600 Mhz
Video Card(s) Sapphire RX 7900 XT - Alphacool Eisblock Aurora
Storage 1x Kingston KC3000 1TB 1x Kingston A2000 1TB, 1x Samsung 850 EVO 250GB , 1x Samsung 860 EVO 500GB
Display(s) LG UltraGear 32GN650-B + 4K Samsung TV
Case Phanteks NV7
Power Supply GPS-750C
There is no IPC improvement if you compare cuda core count.

GPUs are designed differently than CPU , there are little IPC gains that can be had. Most major improvements come from increased core count and clocks.
 
Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
Titan V is the new flagship GPU it will be approximately one par with ~ GTX1180TI look at GPU history! If it's true than there is no place for big performance gain over next generation.
 
Last edited:

ppn

Joined
Aug 18, 2015
Messages
1,231 (0.39/day)
50%, in case of 16Gbps G6, and 30% in case of 12GBps. as for the 70 card we have 40% gpu and 50% for the memory so roughly 45%.
 
Joined
Apr 12, 2013
Messages
6,728 (1.68/day)
According to some rumors the next mainstream dGPU will be GT104 i.e. Turing, not Volta, needless to say YMWV with such stuff. It's quite possibe that Nvidia may finally go the Intel route & seperate their (desktop) gaming & workstation cards at the hip, not necessarily castrate them like they've done in the past.
 
Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
According to some rumors the next mainstream dGPU will be GT104 i.e. Turing, not Volta
Thats is just a name i don't think it will be different architecture they probably just will cut off tenser cores on smaler die maybe some minor improvements. Look at Intel Skylake, Kaby Lake and Coffee Lake basically the same architecture now for how many years ?

Right now also nVIDIA has no competition with AMD so.... they don't really need huge performance increase over previous generation GPUs. It's sad to say but AMD in GPU market is nobody if comparing to nVIDIA in raw power and performance per watt.
 
Last edited:
Joined
Apr 12, 2013
Messages
6,728 (1.68/day)
Thats is just a name i don't think it will be different architecture they probably just will cut off tenser cores on smaler die maybe some minor improvements. Look at Intel Skylake, Kaby Lake and Coffee Lake basically the same architecture now for how many years ?

Right now also nVIDIA has no competition with AMD so.... they don't really need huge performance increase over previous generation GPUs. It's sad to say but AMD in GPU market is nobody if comparing to nVIDIA in raw power and performance per watt.
Skylake & Skylake X aren't the same, that's what I was referrring to when I said that about Nvidia, right now they've enough money (& demand) to pull something similar off.
 
Joined
Jul 19, 2006
Messages
43,585 (6.75/day)
Processor AMD Ryzen 7 7800X3D
Motherboard ASUS TUF x670e
Cooling EK AIO 360. Phantek T30 fans.
Memory 32GB G.Skill 6000Mhz
Video Card(s) Asus RTX 4090
Storage WD m.2
Display(s) LG C2 Evo OLED 42"
Case Lian Li PC 011 Dynamic Evo
Audio Device(s) Topping E70 DAC, SMSL SP200 Headphone Amp.
Power Supply FSP Hydro Ti PRO 1000W
Mouse Razer Basilisk V3 Pro
Keyboard Tester84
Software Windows 11
Performance is dictated by money. With no competition I would expect around 20% more performance on average with the new cards. I don't see why Nvidia should give any more than that.
 
Joined
Aug 6, 2017
Messages
7,412 (3.07/day)
Location
Poland
System Name Purple rain
Processor 10.5 thousand 4.2G 1.1v
Motherboard Zee 490 Aorus Elite
Cooling Noctua D15S
Memory 16GB 4133 CL16-16-16-31 Viper Steel
Video Card(s) RTX 2070 Super Gaming X Trio
Storage SU900 128,8200Pro 1TB,850 Pro 512+256+256,860 Evo 500,XPG950 480, Skyhawk 2TB
Display(s) Acer XB241YU+Dell S2716DG
Case P600S Silent w. Alpenfohn wing boost 3 ARGBT+ fans
Audio Device(s) K612 Pro w. FiiO E10k DAC,W830BT wireless
Power Supply Superflower Leadex Gold 850W
Mouse G903 lightspeed+powerplay,G403 wireless + Steelseries DeX + Roccat rest
Keyboard HyperX Alloy SilverSpeed (w.HyperX wrist rest),Razer Deathstalker
Software Windows 10
Benchmark Scores A LOT
1180 has got to be faster than 1080Ti. Especially considering the 2nd hand market is gonna be flooded with those. If it isn't, I am waiting for 1180Ti. If I can get 1.5x of 1080 performance I won't mind a high pricetag. I think nvidia knows that in the enthusiast segment the higher the performance, the higher the profit they can make. Refreshing the lineup with marginal performance gain is going to stagnate the sales.
 
Last edited:
Joined
Mar 10, 2010
Messages
11,878 (2.32/day)
Location
Manchester uk
System Name RyzenGtEvo/ Asus strix scar II
Processor Amd R5 5900X/ Intel 8750H
Motherboard Crosshair hero8 impact/Asus
Cooling 360EK extreme rad+ 360$EK slim all push, cpu ek suprim Gpu full cover all EK
Memory Corsair Vengeance Rgb pro 3600cas14 16Gb in four sticks./16Gb/16GB
Video Card(s) Powercolour RX7900XT Reference/Rtx 2060
Storage Silicon power 2TB nvme/8Tb external/1Tb samsung Evo nvme 2Tb sata ssd/1Tb nvme
Display(s) Samsung UAE28"850R 4k freesync.dell shiter
Case Lianli 011 dynamic/strix scar2
Audio Device(s) Xfi creative 7.1 on board ,Yamaha dts av setup, corsair void pro headset
Power Supply corsair 1200Hxi/Asus stock
Mouse Roccat Kova/ Logitech G wireless
Keyboard Roccat Aimo 120
VR HMD Oculus rift
Software Win 10 Pro
Benchmark Scores 8726 vega 3dmark timespy/ laptop Timespy 6506
1180 has got to be faster than 1080Ti. Especially considering the 2nd hand market is gonna be flooded with those. If it isn't, I am waiting for 1180Ti. If I can get 1.5x of 1080 performance I won't mind a high pricetag. I think nvidia knows that in the enthusiast segment the higher the performance, the higher the profit they can make. Refreshing the lineup with marginal performance gain is going to stagnate the sales.
I think Nvidia will play it wise, 20-30% and regardless of name their shader cores are unlikely to change, they didn't on the Titan V, the work went into tensor cores and integration.
The 1180 will greatly resemble a 1080 with faster memory on an improved process because that's what it will be.

With a overstocked channel now and a second hand market on the rise as expected, and the best performing card already out, expect high prices for these, with low stock too since it would not make sense to flood the market with the channel full.

At present Nvidia are focused on Ai and data centres not gaming cards and why would they be , they're taking a leaf from intels book , when at the top 10-20% will do and that doesn't require shaders to be re architected, just a process kick up and faster memory, as i said, Turing ,pah ,they are just names and mearly indicate A difference not a significant one though.
 
Joined
Jul 24, 2011
Messages
90 (0.02/day)
Location
phliadelphia
Processor i5 6500 @ 4.5
Motherboard Asus z170
Memory 16gb ddr4 3000
Video Card(s) gtx 1070
Storage 1tb Seagate 7200 rpm
Case Antec 1200
Power Supply Corsair 750
Software windows 10 pro
I'm guessing 40 to 50% faster.
NVIDIA-GeForce-GTX-1170-Relative-Performance-wccftech.png

https://wccftech.com/873853-2/
 
Joined
Aug 13, 2010
Messages
5,369 (1.08/day)
From GTX480 (40nm) to GTX580 (40nm) performance gain ~ 16%
Barely a refresh
From GTX580 (40nm) to GTX680 (28nm) performance gain ~ 30%
Moving from a huge mother-die, to a very small middle of the pack one
From GTX680 (28nm) to GTX780 (28nm) performance gain ~ 21%
Moving from the middle chip to the almost-high end of the same architecture
From GTX780 (28nm) to GTX980 (28nm) performance gain ~ 30%
A move from a large die to again, a small 66%-of-full one
From GTX980 (28nm) to GTX1080 (16nm) performance gain ~ 67%
An actual move from one middle of the pack chip, to a completely new architecture and lithography while keeping the same power consumption.

This comparison is a bit problematic IMO, since prices and positions are a bit different albeit similar monikers.
Full enabled chips should be compered to other fully enabled ones, cut-down by 30-35% ones to cut-down ones. Per-die size and power consumpion, the changes were actually huge.

We're talking GTX 580 to GTX 780 Ti to GTX 980 Ti to GTX 1080 Ti, or GTX 460 to GTX 660 to GTX 960 to GTX 1060. Those are incredible gains.
 
Last edited:
Joined
Dec 6, 2016
Messages
748 (0.28/day)
A realistic view on the situation: The "12nm" process is only a smaller version of the "16nm" process. Unless nVidia decides to do some massive architecture changes, 15-20% improvement at the same power consumption is what we will probably get.
 
Joined
Dec 18, 2005
Messages
8,253 (1.24/day)
System Name money pit..
Processor Intel 9900K 4.8 at 1.152 core voltage minus 0.120 offset
Motherboard Asus rog Strix Z370-F Gaming
Cooling Dark Rock TF air cooler.. Stock vga air coolers with case side fans to help cooling..
Memory 32 gb corsair vengeance 3200
Video Card(s) Palit Gaming Pro OC 2080TI
Storage 150 nvme boot drive partition.. 1T Sandisk sata.. 1T Transend sata.. 1T 970 evo nvme m 2..
Display(s) 27" Asus PG279Q ROG Swift 165Hrz Nvidia G-Sync, IPS.. 2560x1440..
Case Gigabyte mid-tower.. cheap and nothing special..
Audio Device(s) onboard sounds with stereo amp..
Power Supply EVGA 850 watt..
Mouse Logitech G700s
Keyboard Logitech K270
Software Win 10 pro..
Benchmark Scores Firestike 29500.. timepsy 14000..
pure guesswork but i would say %25 to %30 more grunt with maybe a little less power consumption..

as has been said Nvidia have no reason to do more but any less and people wont bother to upgrade.. having said that they are expecting up to a %60 drop in sales from falling mining demand.. maybe they will try harder just to get some of this back..

trog
 
Joined
Sep 17, 2014
Messages
20,692 (5.96/day)
Location
The Washing Machine
Processor i7 8700k 4.6Ghz @ 1.24V
Motherboard AsRock Fatal1ty K6 Z370
Cooling beQuiet! Dark Rock Pro 3
Memory 16GB Corsair Vengeance LPX 3200/C16
Video Card(s) ASRock RX7900XT Phantom Gaming
Storage Samsung 850 EVO 1TB + Samsung 830 256GB + Crucial BX100 250GB + Toshiba 1TB HDD
Display(s) Gigabyte G34QWC (3440x1440)
Case Fractal Design Define R5
Audio Device(s) Harman Kardon AVR137 + 2.1
Power Supply EVGA Supernova G2 750W
Mouse XTRFY M42
Keyboard Lenovo Thinkpad Trackpoint II
Software W10 x64
Titan V is the new flagship GPU it will be approximately one par with ~ GTX1180TI look at GPU history! If it's true than there is no place for big performance gain over next generation.

No, you need to get that out of your head. The flagship GPU right now is the 1080ti. Pascal is the gaming 'state of the art' today. Not Volta.

GV100 / Titan V is not a Geforce segment card, was never marketed as such, only a few deluded nubs and rich kids think 'Titan' has anything to do with gaming. It does NOT. Titan simply is Nvidia's playground to push the envelope in whatever way they think the market is moving. And they have been completely correct so far: when Titan moved out of compute and into 'fastest gaming GPU' the market moved along in that direction with a strong desire for high-end gaming GPUs to drive 1440p and 4K. Today, Titan answers to the strong market push for AI / machine learning. Tomorrow? Who knows, perhaps it'll do your laundry or controls your smart home.

Barely a refresh

Moving from a huge mother-die, to a very small middle of the pack one

Moving from the middle chip to the almost-high end of the same architecture

A move from a large die to again, a small 66%-of-full one

An actual move from one middle of the pack chip, to a completely new architecture and lithography while keeping the same power consumption.

This comparison is a bit problematic IMO, since prices and positions are a bit different albeit similar monikers.
Full enabled chips should be compered to other fully enabled ones, cut-down by 30-35% ones to cut-down ones. Per-die size and power consumpion, the changes were actually huge.

We're talking GTX 580 to GTX 780 Ti to GTX 980 Ti to GTX 1080 Ti, or GTX 460 to GTX 660 to GTX 960 to GTX 1060. Those are incredible gains.

Even this comparison doesn't completely work out, because across all these generations we're also looking at different APIs, different mainstream resolutions, different Shader Model & other feature versions, etc etc etc. Old cards simply lack the efficiency bumps newer generations have not only in hardware but in what code they support well.

Performance is dictated by money. With no competition I would expect around 20% more performance on average with the new cards. I don't see why Nvidia should give any more than that.

Because more performance gives them the opportunity to increase the price and margin. Pascal is a perfect example: it was marketed *with an MSRP increase* across the entire stack of products and we took it for granted exactly because of that reason. Do note: this is the exact 'no competition' situation you're talking about - it is exactly at this time that they made one of the biggest jumps forward.
 
Last edited:
Joined
May 8, 2016
Messages
1,730 (0.60/day)
System Name BOX
Processor Core i7 6950X @ 4,26GHz (1,28V)
Motherboard X99 SOC Champion (BIOS F23c + bifurcation mod)
Cooling Thermalright Venomous-X + 2x Delta 38mm PWM (Push-Pull)
Memory Patriot Viper Steel 4000MHz CL16 4x8GB (@3240MHz CL12.12.12.24 CR2T @ 1,48V)
Video Card(s) Titan V (~1650MHz @ 0.77V, HBM2 1GHz, Forced P2 state [OFF])
Storage WD SN850X 2TB + Samsung EVO 2TB (SATA) + Seagate Exos X20 20TB (4Kn mode)
Display(s) LG 27GP950-B
Case Fractal Design Meshify 2 XL
Audio Device(s) Motu M4 (audio interface) + ATH-A900Z + Behringer C-1
Power Supply Seasonic X-760 (760W)
Mouse Logitech RX-250
Keyboard HP KB-9970
Software Windows 10 Pro x64
No, you need to get that out of your head. The flagship GPU right now is the 1080ti. Pascal is the gaming 'state of the art' today. Not Volta.

GV100 / Titan V is not a Geforce segment card, was never marketed as such, only a few deluded nubs and rich kids think 'Titan' has anything to do with gaming. It does NOT.

Even this comparison doesn't completely work out, because across all these generations we're also looking at different APIs, different mainstream resolutions, different Shader Model & other feature versions, etc etc etc. Old cards simply lack the efficiency bumps newer generations have not only in hardware but in what code they support well.
I think you missing the point here.
"Flagship" is decided by marketing. Chips and cards are designed by engineers.
OP wanted to measure increase in speed/performance across generations and not in naming schemes, or cash grabbing.

Fastest Fermi = GF110 (GTX 580 "3GB")
Fastest Kepler = GK110b (Titan Black)
Fastest Maxwell (2.0) = GM200 (Titan X "M")
Fastest Pascal = GP102/GP100 (Titan Xp)

Fastest card NV can do in Pascal design is Titan Xp, not GTX 1080 Ti.
I refuse to call latter one a "flagship" because fagships should represent the best company can do at given point in time.
If NV wanted too, they could release fully unlocked 3840 CC GP102 (with GDDR5X memory, because HBM 2.0 will be limited to Titan/Quadro/Tesla series), as for example "Geforce GTX 1090".
They have the means, it's simply not economic to do that at this point.
 
Joined
Sep 17, 2014
Messages
20,692 (5.96/day)
Location
The Washing Machine
Processor i7 8700k 4.6Ghz @ 1.24V
Motherboard AsRock Fatal1ty K6 Z370
Cooling beQuiet! Dark Rock Pro 3
Memory 16GB Corsair Vengeance LPX 3200/C16
Video Card(s) ASRock RX7900XT Phantom Gaming
Storage Samsung 850 EVO 1TB + Samsung 830 256GB + Crucial BX100 250GB + Toshiba 1TB HDD
Display(s) Gigabyte G34QWC (3440x1440)
Case Fractal Design Define R5
Audio Device(s) Harman Kardon AVR137 + 2.1
Power Supply EVGA Supernova G2 750W
Mouse XTRFY M42
Keyboard Lenovo Thinkpad Trackpoint II
Software W10 x64
I think you missing the point here.
"Flagship" is decided by marketing. Chips and cards are designed by engineers.
OP wanted to measure increase in speed/performance across generations and not in naming schemes, or cash grabbing.

Fastest Fermi = GF110 (GTX 580 "3GB")
Fastest Kepler = GK110b (Titan Black)
Fastest Maxwell (2.0) = GM200 (Titan X "M")
Fastest Pascal = GP102/GP100 (Titan Xp)

Fastest card NV can do in Pascal design is Titan Xp, not GTX 1080 Ti.
I refuse to call latter one a "flagship" because fagships should represent the best company can do at given point in time.
If NV wanted too, they could release fully unlocked 3840 CC GP102 (with GDDR5X memory, because HBM 2.0 will be limited to Titan/Quadro/Tesla series), as for example "Geforce GTX 1090".
They have the means, it's simply not economic to do that at this point.

There is what, 10% performance between the XP and the 1080ti. If you want it, you can have that one... XP is so relevant that I simply forgot about it ;) It does not change the reason Titan exists - its Nvidia's playground and they do whatever they like with it. Remember Titan Z, no one ever considered that a serious option in the gaming space. And Titan V is of the same kind, it does not belong in these comparisons. Even for their completely twisted price/perf metrics the Titans should be ignored in these comparisons. If they really are needed as a gaming GPU in the regular stack, Nvidia releases it as such: enter GTX 780/ti, GTX 580, etc etc. The flagship tells you nothing about relative performance between generations either because its not a regularly appearing product with a singular reason of existance.

The bottom line of whichever approach you choose for any such comparison is going to be that the comparison is flawed. You cannot compare across generations like this without proper context and a boatload of 'ifs and buts'.
 
Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
Hopefully when GTX 1180 will come out in july GTX 1080 Ti prices will drop the same as it happened to GTX 1080 when it was realased right after month GTX 980 Ti price was dropped from ~ 600eur to ~ 430eur.

At that time GTX 980 Ti was better bang for the buck than GTX 1070 new from store.
 
Joined
May 29, 2017
Messages
283 (0.11/day)
Location
Latvia
Processor AMD Ryzen™ 7 5700X
Motherboard ASRock B450M Pro4-F R2.0
Cooling Arctic Freezer A35
Memory Lexar Thor 32GB 3733Mhz CL16
Video Card(s) PURE AMD Radeon™ RX 7800 XT 16GB
Storage Lexar NM790 2TB + Lexar NS100 2TB
Display(s) HP X34 UltraWide IPS 165Hz
Case XPG Starker AIR
Audio Device(s) Airpulse SM200 + Volt 1 + Rode NT1 SS
Power Supply Cougar GEX 750
Mouse Cooler Master MM730
Keyboard Krux Atax PRO Gateron Yellow
Software Windows 10 Pro
Thats is just a name i don't think it will be different architecture they probably just will cut off tenser cores on smaler die maybe some minor improvements.

NVIDIA's upcoming GeForce GTX graphics cards based on the "Volta" architecture, could feature GDDR6 memory, according to a supply deal SK Hynix struck with NVIDIA, resulting in the Korean memory manufacturer's stock price surging by 6 percent. It's not known if GDDR6 will be deployed on all SKUs, or if like GDDR5X, it will be exclusive to a handful high-end SKUs. The latest version of SK Hynix memory catalogue points to an 8 Gb (1 GB) GDDR6 memory chip supporting speeds of up to 14 Gbps at 1.35V, and up to 12 Gbps at 1.25V.
 
Joined
Apr 12, 2013
Messages
6,728 (1.68/day)
NVIDIA's upcoming GeForce GTX graphics cards based on the "Volta" architecture, could feature GDDR6 memory, according to a supply deal SK Hynix struck with NVIDIA, resulting in the Korean memory manufacturer's stock price surging by 6 percent. It's not known if GDDR6 will be deployed on all SKUs, or if like GDDR5X, it will be exclusive to a handful high-end SKUs. The latest version of SK Hynix memory catalogue points to an 8 Gb (1 GB) GDDR6 memory chip supporting speeds of up to 14 Gbps at 1.35V, and up to 12 Gbps at 1.25V.
Names aside, it doesn't matter too much what Nvidia releases, it should still be a good jump from Pascal.
  • Updated AMD Zen 2, Zen+ second wave, Threadripper 2nd Gen
  • Added GDDR6 info for NVIDIA
  • Added AMD B450 chipset
  • Added QLC NAND flash
  • Added NVIDIA GTX 1050 3 GB
  • Updated NVIDIA Turing
  • Updated Intel Cannon Lake
  • Updated Intel Z390
  • Updated DDR5
 
Top