MSI GeForce RTX 5070 Ti Ventus 3X OC

[Baba]

I'm not willing to upgrade for less than a 50% performance uplift,

That thinking is why I'm still using my RX6950XT. 7900XTX was only a 50% uplift. Now, I'm stuck until either 5080 or 5090 prices stabilize. I might just get a Switch 2 when it comes out.

 

[razaron]

Oh look at that all of the ones for 749 are gone on Newegg.. lol but if you really want one then you can get one off of ebay in a few minutes...

EDIT: It's early.. did they ever show instock because they were always showing out of stock.

There was 1 on ebuyer for £729, it went out of stock so fast their frontend couldn't keep up. Clicking buy showed 1 item in cart, click cart to see cart empty lol

 

[Makaveli]

That thinking is why I'm still using my RX6950XT. 7900XTX was only a 50% uplift. Now, I'm stuck until either 5080 or 5090 prices stabilize. I might just get a Switch 2 when it comes out.

6800XT to 7900XTX was 50% i did this one in 2023.

6950XT to 7900XTX about 35%.

I also generally don't move up in gpu's until I hit that 50% sweet spot also.

 

[JustBenching]

Your whole take on this is quite literally retarded. Instead of blaming crap developers for making s**tty PC ports of console games, you blame W1zz for blaming those developers. I cannot stress enough how mind-bogglingly braindead your view is; I've come across some dumb takes in my 2 decades online, but this one is up there with the stupidest.

It is not NVIDIA's fault that developers make s**tty ports, nor is NVIDIA required - and nor should they be - to build their hardware to cater for s**tty ports.

Stop posting dumb crap and go sit in the dunce corner and for once THINK about what you're posting, BEFORE you post it.

Especially the game he mentioned as and example, hogwarts. That thing stutters on an overclocked tuned 9800x 3d and a 4090, but damn you nvidia...

 

[Daven]

lol i'm just messing with you people. 7 more minutes until stores open

I like constantly changing headlines. Keeps us on our toes!

 

[Chrispy_]

I managed to snag this exact model for the £729 MSRP at lunch by mashing F5 across 29 bookmarked tabs.

The scalping is real, almost all the "MSRP" cards were £800 or more, many UK retailers adding a scalper tax for the impatient. The next best price was Overclockers.co.uk selling this Ventus for £800 so a 10% scalping markup direct from an official Nvidia reselling partner (shop.nvidia.com/en-gb links directly to OCUK)

 

[Why_Me]

I managed to snag this exact model for the £729 MSRP at lunch by mashing F5 across 29 bookmarked tabs.

The scalping is real, almost all the "MSRP" cards were £800 or more, many UK retailers adding a scalper tax for the impatient. The next best price was Overclockers.co.uk selling this Ventus for £800 so a 10% scalping markup direct from an official Nvidia reselling partner (shop.nvidia.com/en-gb links directly to OCUK)

The Brits have no idea how good they have it.

https://uk.pcpartpicker.com/products/video-card/#c=589

 

[alwayssts]

6800XT to 7900XTX was 50% i did this one in 2023.

6950XT to 7900XTX about 35%.

I also generally don't move up in gpu's until I hit that 50% sweet spot also.

I think a better way to look at it for many is going forward (with all ML upscalers), whenever your current card can't do 1440p at the settings you want (be it upscaled from 960p, native, or 1440p->4k), upgrade.
4k is just asking too much; heck asking 1440pRT is asking too much. Likewise, many are not going to happy with lower. 1440p native, 1080p->4k, or 960->1440p all make sense to varying degrees.

DLSS3 had generally a ~8% perf hit in 'quality'. DLSS4 15%. FSR3 between the two.
FSR4 IDK but I think it'll be ~8-9% (but this could be made up through the rop limitation/shader cache, long-term UDNA goal?). This is kind of an important variable.
RT has a roughly 31% performance hit on nVIDIA, and it will likely be the same on AMD's newer arch, as I think the extra TMUs will effectively do the same thing as nVIDIA's fixed-function hw (30%+ from Navi3).

Does that help anyone? IDK. But there ya go. Situation is fluid, but in the end I think your tiers would essentially be 9070xt/4070ti, 5080 24GB, and 4090. One of those cards doesn't exist.

These will probably be replaced by 6144sp, 9216, and 12288sp with 16/18/24GB of ram on 3nm and perform more-or-less the same. Good, better, best. Not really gonna change, just slip one tier down.

I think asking $700 for the bottom tier is ridiculous, but I understand it relatively speaking. For 5080 to be worth $1000, it needs to have 24GB. 4090 prolly not too far off MSRP; all perfectly relative.
These will all likely drop one price tier next-gen, and I'm sure there will be some absurd tier above them that almost nobody buys, probably. But it's pretty much that simple.

IMO these prices could still all be relative and should be priced one tier down already, but they are not because they are propping up next-gen prices. Whaddya do, that's life ig. DNW.

 

[Makaveli]

I think a better way to look at it for many is going forward (with all ML upscalers), whenever your current card can't do 1440p at the settings you want (be it upscaled from 960p, native, or 1440p->4k), upgrade.
4k is just asking too much; heck asking 1440pRT is asking too much. Likewise, many are not going to happy with lower. 1440p native, 1080p->4k, or 960->1440p all make sense to varying degrees.

DLSS3 had generally a ~8% perf hit in 'quality'. DLSS4 15%. FSR3 between the two.
FSR4 IDK but I think it'll be ~8-9% (but this could be made up through the rop limitation/shader cache, long-term UDNA goal?). This is kind of an important variable.
RT has a roughly 31% performance hit on nVIDIA, and it will likely be the same on AMD's newer arch, as I think the extra TMUs will effectively do the same thing as nVIDIA's fixed-function hw (30%+ from Navi3).

Does that help anyone? IDK. But there ya go. Situation is fluid, but in the end I think your tiers would essentially be 9070xt/4070ti, 5080 24GB, and 4090. One of those cards doesn't exist.

These will probably be replaced by 6144sp, 9216, and 12288sp with 16/18/24GB of ram on 3nm and perform more-or-less the same. Good, better, best. Not really gonna change, just slip one tier down.

I think asking $700 for the bottom tier is ridiculous, but I understand it relatively speaking. For 5080 to be worth $1000, it needs to have 24GB. 4090 prolly not too far off MSRP; all perfectly relative.
These will all likely drop one price tier next-gen, and I'm sure there will be some absurd tier above them that almost nobody buys, probably. But it's pretty much that simple.

IMO these prices could still all be relative and should be priced one tier down already, but they are not because they are propping up next-gen prices. Whaddya do, that's life ig. DNW.

Valid point but i'm someone that doesn't use upscaling or Frame gen at all.

So for me it will be 50% at native.

And I agree all tiers are over priced at the moment. Ever since the first bitcoin mining event vendors have been taking advantage and keeping prices high. After what happen with EVGA and NV I understand it alittle because those margins are super thin just sucks that we as customers feel the brunt of it.

 

[Chrispy_]

The Brits have no idea how good they have it.

https://uk.pcpartpicker.com/products/video-card/#c=589

Nah, that's just PCPartPicker being out of date. There's no stock for less than £800 now ($1011, €966)

 

[Assimilator]

Perhaps he's confusing the VRAM usage that gets listed on individual game reviews such as https://www.techpowerup.com/review/kingdom-come-deliverance-ii-performance-benchmark/

Or perhaps he's just as thick as his posts show. Hanlon's razor, and all that.

 

[AnotherReader]

N3P brings +60% logic density, therefore we're looking at 7680->12288C 18GB for the 6070/Ti, 6080 with 16384C 24GB. Rubin is released earlier, like Pascal, so there's no need for a super on N4. What for.

GPUs have large amounts of SRAM in their registers and caches as well; that won't scale at all with N3. In addition, PHYs for memory and PCIe won't scale either.

In Canada, this one is being sold for the equivalent of 880 USD. It seems the MSRP, at least for this SKU, is illusory.

 

[Makaveli]

GPUs have large amounts of SRAM in their registers and caches as well; that won't scale at all with N3. In addition, PHYs for memory and PCIe won't scale either.

In Canada, this one is being sold for the equivalent of 880 USD. It seems the MSRP, at least for this SKU, is illusory.

View attachment 385779

The asus model's aren't much better. And those prime models probably have worse coolers than the TUF model.

 

[Dawora]

• Good performance gains: +30% over 4070 Ti
• WHAT !?!?!?!?
• ))))))))))))))))))))))))))))))))))))))))))))))))
• This the best joke in the last 2 mounts !

Better than many ppls think about
Ppls told in many forums it will be +10% vs 4070Ti
So ofc 30% is good and thats how it should be..

But whiners whine, what ever hapens, its need to be free and +100% faster than 4090 to not need to read whiners postings in forums, and still im sure some ppls whine..

 

[lexluthermiester]

That's nonsense. Why is the inflation only affecting GPUs, while new CPUs are released at lower MSRPs than older gens? It's the same manufacturing process.

No, it is NOT nonsense. Inflation has been big problem for everyone, everywhere in this post-pandemic time period. It has affected food stuffs, household goods, automotive etc, etc, etc. A package of burrito's before the pandemic was $2.50. That exact same package now costs $5.20. That is more than double and is just ONE example. I could keep going like this for days.

Inflation is a factor here. This can not be denied.

I'm not trying to justify anyone's greed, however the rise of prices from 5 years ago is EASILY explained. Anyone who fails to recognize simple fact and reality is a fool unto themselves and competes with tungsten where density is concerned.

Having said that, the following is not inflation, it is blatant, disgusting greed;

The asus model's aren't much better. And those prime models probably have worse coolers than the TUF model.

View attachment 385792

@ ASUS and anyone else price gouging;
You are morons. Pathetic losers. No money for you until you drop your prices to MSRP or near-MSRP levels. Otherwise, go boil your heads.

 

[N/A]

GPUs have large amounts of SRAM in their registers and caches as well; that won't scale at all with N3. In addition, PHYs for memory and PCIe won't scale either.

In Canada, this one is being sold for the equivalent of 880 USD. It seems the MSRP, at least for this SKU, is illusory.

Precisely Sram and phy remain untouched at 48/192b, 64/256b and 96MB/384bit where 18432 Cuda was previously taking up to 252mm2 on N3E becomes 30720 Cuda or 24576 for a more conservative number of enabled logic. So you'll get the generational uplift then hopefully with faster mem and 24Gbit chips, this gen is just a mirage anyway.

 

[Hecate91]

Perhaps he's confusing the VRAM usage that gets listed on individual game reviews such as https://www.techpowerup.com/review/kingdom-come-deliverance-ii-performance-benchmark/

That's what I was thinking of, thank you. Since VRAM usage gets listed for individual game reviews, I don't see why not for the whole suite of games, or at minimum the more VRAM heavy titles.

 

[AnotherReader]

Precisely Sram and phy remain untouched at 48/192b, 64/256b and 96MB/384bit where 18432 Cuda was previously taking up to 252mm2 on N3E becomes 30720 Cuda or 24576 for a more conservative number of enabled logic. So you'll get the generational uplift then hopefully with faster mem and 24Gbit chips, this gen is just a mirage anyway.

Each SMX has significant amounts of SRAM that your calculation is neglecting: 128 KB of L1 and 256 KB of registers in each Ada SM. Therefore, the SMX won't scale as well as you think.

 

[chrcoluk]

The Brits have no idea how good they have it.

https://uk.pcpartpicker.com/products/video-card/#c=589

We are one of the few FE countries which does help on Nvidia as FE is guaranteed MSRP, but we not a AMD country for their equivalent product.

 

[GodisanAtheist]

Nah, at least you have a chance there to get one and at MSRP, as soon as you do what you suggest, bots will buy them all up.

- I don't understand why folks are having a hard time understanding what I said.

Great, I love that MC is there for folks who live nearby and are able to physically go to a store and snag a great deal. It's great that they have that model and they should stick with it.

THAT SAID

People need to stop using MC pricing to win internet arguments about pricing. Yes, there is a posted price there, but it is irrelevant to a huge number of people even here in the US thanks to MC's very limited footprint and unwillingness to offer those prices for online purchases (which would make them relevant to the entire US market).

It's like telling people eggs are free because you have a chicken coup in the backyard while they struggle to pay for groceries due to inflation. It's just unproductive talking passed each other.

 

[lexluthermiester]

People need to stop using MC pricing to win internet arguments about pricing.

Ok, BestBuy then. There's almost no difference.

It's like telling people eggs are free because you have a chicken coup in the backyard while they struggle to pay for groceries due to inflation. It's just unproductive talking passed each other.

True.

 

[GodisanAtheist]

Ok, BestBuy then. There's almost no difference.


True.

- I'd be fine with that, they're a nationwide electronics chain with locations in all 50 states (at least the continuous 48) with the same instore and online ordering prices.

 

[AnotherReader]

- I'd be fine with that, they're a nationwide electronics chain with locations in all 50 states (at least the continuous 48) with the same instore and online ordering prices.

They have a significant presence in Canada too: 333 stores.

 

[Knight47]

The asus model's aren't much better. And those prime models probably have worse coolers than the TUF model.

The Prime cooler is better than the Ventus, Ventus has 4 heatpipes, Prime has 6. But nowhere near TUF.
This launch was worse here than 5080, we had a few 5070ti cards for almost the same price as the 5080. Looking at the 5070 specs, that will be slower than the 4070Super and 5060Ti will barely beat the 4060Ti.

 

[lexluthermiester]

The Prime cooler is better than the Ventus, Ventus has 4 heatpipes, Prime has 6. But nowhere near TUF.
This launch was worse here than 5080. Looking at the 5070 specs, that will be slower than the 4070Super and 5060Ti will barely beat the 4060Ti.

Looking at the temps, the two extra heatpipes don't really matter much.