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

I interpret the 5000 series like this.

Joined
Apr 13, 2017
Messages
303 (0.10/day)
System Name nVAMDia
Processor Ryzen 7900 at 145Watts PPT.
Motherboard MSI MAG x670 Tomahawk Wifi
Cooling AIO240 for CPU, Wraith Prism's Fan for RAM but suspended above it without touching anything in case.
Memory 32GB dual channel Gskill DDR6000CL30
Video Card(s) Zotac 5070 gaming solid oc + Msi Ventus 2x Rtx 4070
Storage Samsung Evo 970
Display(s) Old 1080p 60FPS Samsung
Case Normal atx
Audio Device(s) Dunno
Power Supply 1200Watts (8x 8-pin)
Mouse wireless & quiet
Keyboard wireless & quiet
VR HMD No
Software Windows 11
Benchmark Scores The best GPU in the world is the brain. 100!
Rtx5090: actually a titan.

Rtx5080: gap is similar to gtx670 vs gtx titan so this is actually a 5070.

Rtx5070: has very close memory characteristics to 5080. So its not gtx660ti. Its just a cropped 5070 (due to production fault) so its cheaper.

Rtx5060 ti: with more memory but slower memory than 5070, it looks like a gtx 660 ti 3GB version. Its truly a 60ti segment. If price is right, could be the sweet spot for the most popular resolution that is 1080p.

Rtx5060: half the memory, low bandwidth, the low end that is still way better in low settings, I think people will not buy this much. Except when they have no space in pc for a bulky heatspreader.
 
might as well go radeon then.
 
Rx9070xt looks like an Hd7970 GHz edition.
 
We need the prices to make a real judgement, but yeah, hard to imagine the 5090 being less than 2.5k given it's specs, that is Titan(ic?) pricing.
 
5090 could even be 2x 5080 "glued"??
 
No one gives a flying fuck about the 1% of people who buy **90 series, yes, prosumers, content creators, AI etc etc maybe stop being a cheap schmuck and buy a professional card for your needs and leave the gamers to have decent GPU's at a price that doesn't require a remortgage or selling a feckin kidney, you want professional performance then pony up and buy a pro card, getting sick of mid range 60 series cards costing upto $500 cause of the few idiots who don't give a shit about anyone else, also why NV can get bent as I won't buy into their Apple style closed ecosystem and pay a 200% premium for it, so yes, I will stick to AMD or even maybe Intel, would rather piss my money away than give it to NV who have grown way too big for their boots and think they are untouchable, the downfall will be immense! still the sheep will watch YT influencers and go with the narrative, you can't stop the herd mentality from the sheep unfortunately

I'll let you know as soon as I get one.
Imagine at laughing at the average man in the street who can't afford 2k GPU's and just want to enjoy their hobby whilst they have some downtime from work, family, life commitments etc to relax, just because you are driving this machine and letting companies price everyday normal people out of the market and paying them every gen, and getting joy out of it....
 
No one gives a flying fuck about the 1% of people who buy **90 series, yes, prosumers, content creators, AI etc etc maybe stop being a cheap schmuck and buy a professional card for your needs and leave the gamers to have decent GPU's at a price that doesn't require a remortgage or selling a feckin kidney, you want professional performance then pony up and buy a pro card, getting sick of mid range 60 series cards costing upto $500 cause of the few idiots who don't give a shit about anyone else, also why NV can get bent as I won't buy into their Apple style closed ecosystem and pay a 200% premium for it, so yes, I will stick to AMD or even maybe Intel, would rather piss my money away than give it to NV who have grown way too big for their boots and think they are untouchable, the downfall will be immense! still the sheep will watch YT influencers and go with the narrative, you can't stop the herd mentality from the sheep unfortunately
Spoken like a true person who gives that gives that FF because I saved my nickels and dimes for the next card but didn't sell a kidney or take on a mortgage BS for a video card or any other part you can think of. Some of you kids are silly sounding souls. I'm an old man, retired, made some good choices and some people like you would call me a double-dipper. So I'm doing fine for the moment.
My money is ready and waiting. I got 3 credit cards + BestBuy with ZERO balance. Computer parts ain't squat
 
Rtx5090: actually a titan.

If you follow a trend of each successive gen after GTX 1xxx meaningfully subtracting 10 from the name. This would be a welcome announcement of the flagship RTX 50/50.

Halfway to $5000 and half as powerful as it should be. Negative expectations might open a larger delta yet this gen.
 
Try to keep a level head guys.
 
Keep going, want more points? I got more to give.

Keep em coming.
 
Rtx5090: actually a titan.
The 5090 is not a TITAN at all, a TITAN GPU would have a full die enabled aka GB202 (192SM) in this case and 2x VRAM aka 64GB (32GB per PCB side).
 
Rtx5090: actually a titan.

Rtx5080: gap is similar to gtx670 vs gtx titan so this is actually a 5070.

Rtx5070: has very close memory characteristics to 5080. So its not gtx660ti. Its just a cropped 5070 (due to production fault) so its cheaper.

Rtx5060 ti: with more memory but slower memory than 5070, it looks like a gtx 660 ti 3GB version. Its truly a 60ti segment. If price is right, could be the sweet spot for the most popular resolution that is 1080p.

Rtx5060: half the memory, low bandwidth, the low end that is still way better in low settings, I think people will not buy this much. Except when they have no space in pc for a bulky heatspreader.
RTX 5090, I agree on calling it a titan, like the 4090 its out of reach on price for most gamers, a card with only a gaming name, businesses buy these up, then nvidia can claim higher profits in the gaming sector.
5080- I'll be a bit generous and call it a 5070Ti, its half of what the 5090 will be.
5070, too cut down, should be the 5060Ti.
5060Ti, since it has slower memory it should be called the 5060.
5060, should be the 5050, people will buy it anyway because its nvidia, it has brand prestige and gamers will think its better because they see the halo card and only want nvidia.
 
Rx9070xt looks like an Hd7970 GHz edition.

If the 5080 is analogous to a 670 in positioning, and the 9070 XT is analogous to the 7970 GE... In this circa ~2012 analogy that would make the 9070 XT faster than an RTX 5080? Because the 7970 GHz Edition was created to equal the GTX 680 at launch, and matured to roughly 10% faster by the time it was EOL. Current rumors suggest the 9070 XT to be essentially equal to the outgoing 7900 XT, but that card is slower than a 4080 SUPER, so the 5080 is slower than the existing 4080 SUPER.

This is a really contrived and confusing analogy.
 
RTX 5090, I agree on calling it a titan, like the 4090 its out of reach on price for most gamers, a card with only a gaming name, businesses buy these up, then nvidia can claim higher profits in the gaming sector.
The only reason why Professionals are buying 4090s is because the "Quadro" RTX 6000 Ada with 48GB VRAM is $6800 aka 4.25x the price of a 4090...and their A.I. chips are selling between $30K and $40K!
If Nvidia were not that greedy we would not be in that situation.
 
The only reason why Professionals are buying 4090s is because the "Quadro" RTX 6000 Ada with 48GB VRAM is $6800 aka 4.25x the price of a 4090...and their A.I. chips are selling between $30K and $40K!
If Nvidia were not that greedy we would not be in that situation.
So the best thing we can do as a consumer is vote with our wallet and not sink that much dough into anything that may last us at most what 2-3 years now?
 
Only a few more days before CES, I hope RTX 5000 will be interesting in term of performance leap vs last gen and new features.

As for prices, people can vote with their wallets
 
So the best thing we can do as a consumer is vote with our wallet and not sink that much dough into anything that may last us at most what 2-3 years now?
As an ex flagship user.. :laugh:

For long term use, flagships are the way to fly I think.

Just looking back at my GTX8800, GTX295, GTX285, GTX580, those cards got some good life.

But then you have 1080Ti that broke the modern mold until RTX4090.. almost called it a 4890 because wth Nvidia.

30 and 40 series for me was in the midrange.. so no comment there :D
 
5090 is more like a ~750mm² 2080 Ti with 6% disabled *02 GPU except the 352 bit 11 GB, the former gets full 32GB 512 bit bus but there were rumors for 28 GB.
and 5080 is very similar to 2060 Super with 44% Cuda and 67% Rops enabled and 50% memory config copmpared to a full GB202. SO 5080 is even worse with 44% compared to 47%.
 
Last edited:
5090 is more like a ~750mm² 2080 Ti with 6% disabled *02 GPU except the 352 bit 11 GB, the former gets full 32GB 512 bit bus but there were rumors for 28 GB.
and 5080 is very similar to 2060 Super with 44% Cuda and 67% Rops enabled and 50% memory config copmpared to a full GB202. SO 5080 is even worse with 44% compared to 47%.

We don't now the size of GB203 yet (could be >500mm2), GB203 could have a lot more L2 cache that can feed the cores efficiently.
GB202 could be reserving more die size for more cores instead of more L2 cache, for whatever reasons that only Nvidia engineers know best.
 
  • Like
Reactions: N/A
As an ex flagship user.. :laugh:

For long term use, flagships are the way to fly I think.

Just looking back at my GTX8800, GTX295, GTX285, GTX580, those cards got some good life.

But then you have 1080Ti that broke the modern mold until RTX4090.. almost called it a 4890 because wth Nvidia.

30 and 40 series for me was in the midrange.. so no comment there :D
Considering when I got a R9 290, I got the 290 Vapor X which was a dream card then, it wasn't the max end but neither was the x1950 pro 512 AGP I had before that, which that Skt A system with that gpu was a heavy puncher inspite Athlon 64 and core2 being the top ends. The last top card I had was the Radeon 9700 Pro All In Wonder, I even had a Arctic Cooling VGA Silencer rev3 on it, which Hightech Information Systems Digital started using on their own parts (Arctic Cooling products)

For me I had to spend some dough on an engine because I wasn't going to buy another vehicle and finance it for 5- 7 years with how poorly the global economy is. That was $3 Grand right out of my pocket in cash. My ordeal will be over this year, I will recover some of that loss then go pickup a card that will be in my price range, do some testing with it, and maybe share my results, then Migrate over to a Epyc AM5 or a Threadripper, who knows
 
I just see it as
5090 = Best
5080 = Ok
5070 = Meh
5060 = internet cafe trash
 
Rtx5090: actually a titan.

Rtx5080: gap is similar to gtx670 vs gtx titan so this is actually a 5070.

Rtx5070: has very close memory characteristics to 5080. So its not gtx660ti. Its just a cropped 5070 (due to production fault) so its cheaper.

Rtx5060 ti: with more memory but slower memory than 5070, it looks like a gtx 660 ti 3GB version. Its truly a 60ti segment. If price is right, could be the sweet spot for the most popular resolution that is 1080p.

Rtx5060: half the memory, low bandwidth, the low end that is still way better in low settings, I think people will not buy this much. Except when they have no space in pc for a bulky heatspreader.
I don't think its fair to call the 5080 a 5070. Now I will totally agree we have never seen such a huge gap between the flagship and the 80 card before, but my reasoning is this:

Prior generations, the flagship GeForce cards would be on the biggest die. Flagships being things like the 980 Ti, 1080 Ti, 2080 Ti. Add Titans to the mix such as your Titan Xp's and Titan RTX. With the exception of the 3080 (which was primarily due to using Samsung 8nm I'd wager), and the Kepler 2.0 with the 780, the base 80 non-Ti branded cards have been the step down silicon since the GTX 680. For example, if the 1080 Ti is using GP102, the GTX 1080 uses GP104, etc. In that respect, the 4080 and the 5080 are absolutely no different. AD103 and the upcoming GB203 is still the step down die from the flagship die, AD102 and GB202 respectively.

From there, what has absolutely changed, is the 70 branded tier cards. Going all the way back to GeForce 200-series in 2008, the cards branded as the x70 (or x75 in that case) was a cutdown of the same die used in the 80-card. The GTX 580 used GF110...so did the GTX 570. The GTX 980 and GTX 970...both GM204. Then Turing happened and suddenly you had the flagship 2080 Ti on TU102, the RTX 2080 on TU104...ok this looks familiar. But then RTX 2070? TU106. And to add insult to injury, a price hike! They fixed that with the 2070 Super going to a cutdown TU104 and seems like how its been before right? But then you get Ampere. 90 returns as a product model and effectively replaces both the 80-Ti and Titan tier cards in branding. Sure a 3080 Ti exists, but nobody would call that the flagship of the 30-series. 30-series is just kind of an odd man out having so many product models based on GA102. The 3080 surprisingly enough uses the same die as the 3090, but cutdown. But the 3070? Now that's on GA104...again a different die than the 80 card on GA102.

Fast forward to Lovelace and we see an almost Turing like repeat of silicon tiering and branding with another price hike but it gets a little more ridiculous. Flagship is the 4090 on AD102. 4080 is the step down die, weirdly named AD103 instead of AD104 like previous precedent...strange, but still in line. OK. Then you get 4070 and 4070 Ti, both AD104. What's weird here is the launch of a 4070 Ti at all that Nvidia was first going to try to get away with calling a 4080 as well. But here's where it gets crazy...In the past a 70 Ti has usually been a refresh card or something released later in the generation after AMD releases something. Not this time. The 4070 Ti now takes the place of what really should be the 4070, while the 4070 takes the place of what really should be the 4060 Ti. And they both get a price hike to boot.

So now with the 50-series you are basically getting a repeat, minus the rumor the 5070 Ti will be a cutdown of the 5080 die, so in a sense going back to the previous silicon tiering that existed since the 200-series where 80 and 70 share the same die, but otherwise you likely retain the same price hikes and its worth noting the difference between the flagship and the 80 is bigger than its ever been. But otherwise it's the same branding shenanigans as the 40-series.

In my mind what has happened is this:

xx90 has replaced both Titan and what used to be previously branded xx80 Ti cards that were the flagships of their respective generations.
xx80 with the exception of the 780 and 3080, remains similar to how its been since the 600-series...the step down from the flagship.
xx70 Ti now occupies the space that really would have just been a base 70-non Ti branded card in the past.
xx70 now occupies the space that should really be the xx60 Ti tier.
xx60 Ti now occupies what should have just been the xx60 non-Ti tier.
xx60 is now what was the xx50 Ti or xx50 tier.

So 70 and below branded cards functionally fell in tier of silicon while at the same time getting a price hike. Flagship tier and 80 functionally sit at the same tiering they always have since 600-series, but massively price hiked as well.
 
Back
Top