NVIDIA Readies GeForce RTX 4070 GDDR6 Variant?

[btarunr]

NVIDIA is possibly launching a more cost-effective variant of its GeForce RTX 4070 graphics card featuring GDDR6 memory, replacing the GDDR6X that it originally launched with. The new SKU would be better differentiated from the RTX 4070 SUPER. When NVIDIA refreshed its RTX 40-series "Ada" product stack in January, it discontinued the RTX 4070 Ti and RTX 4080, which had been replaced in the lineup by the RTX 4070 Ti SUPER and RTX 4080 SUPER, but at the time it didn't tinker with the RTX 4070, which continued to sell at a roughly $50 lower price than the RTX 4070 SUPER. This new SKU could be an attempt by NVIDIA to get further down below the $500-mark in pricing.

The RTX 4070 originally launched with 21 Gbps GDDR6X memory. This new variant sees the memory replaced with 20 Gbps conventional GDDR6. The JEDEC standard GDDR6 chips could be cheaper than GDDR6X, and could very well be the same GDDR6 chip models AMD uses in some of its higher Radeon RX 7000 series SKUs. This, however, comes with a 4.75% drop in memory bandwidth, which NVIDIA probably overcomes with increasing the GPU clocks a touch. The ASIC code for this SKU is AD104-251, compared to the AD104-250 of the original RTX 4070. The core configuration is otherwise unchanged—you get 5,888 CUDA cores across 46 streaming multiprocessors. Galax has a card based on this SKU ready.



View at TechPowerUp Main Site | Source

 

[dj-electric]

This can be fair if done in a consumer respectful way.

NVIDIA, you think you can price this fairly cheaper while also giving it a distinctive name in order not to confuse consumers with the GDDR6X model?

NVIDIA, you there?

Hello.... ?

 

[Prima.Vera]

 

[Quicks]

If it was 256bit and 16GB Vram. I would have considered it. But let's screw the customer even more? Nah, Nvidia can go get screwed!

 

[N/A]

299 and we have a deal.

 

[chrcoluk]

Slamming in an extra SKU so they dont have to sell these parts as a lower model.

 

[Vayra86]

x70 and x80 soiled forever with the weirdest stack ever. They only needed one generation to do it. Well done Nvidia.

 

[CyberPomPom]

"better differentiated from the RTX 4070 Super" that's one way to say it..
One could also word it like a silent downgrade at consumer's expense.

 

[Dragokar]

They should at least call it 4070 G6 or so...but the same name is totally stupid.

 

[dont whant to set it"']

Yawn, another 4060 in the lineup.

 

[JWNoctis]

Actual impact is probably minor, but imagine it must be annoying to purchase and only then notice the VRAM difference in small prints, after you plug it in.

"better differentiated from the RTX 4070 Super" that's one way to say it..
One could also word it like a silent downgrade at consumer's expense.

They should at least call it 4070 G6 or so...but the same name is totally stupid.

At least, it's not trying to pass off the RTX 4070 Ti as "RTX 4080 12GB," as they tried to do.

 

[64K]

They should at least call it 4070 G6 or so...but the same name is totally stupid.

Not like they haven't done it before. The 1060 6GB and the 1060 3GB. Not only the VRAM difference but the 6GB version had 11% more cores as well. For a while Minimum System Requirements would just say 1060 and leave it to the gamer to guess which one it was referring to until reviews came out and then there was also the 1060 with faster VRAM.

 

[Event Horizon]

This kind of deception should be illegal.

 

[DaemonForce]

I could see a downstep in product name like 4070D or 4070 Turbo.
No idea what impact there would be on performance but GDDR6X was already chosen first for good reason.
It doesn't make sense to revisit the product stack and bin new stuff when there's already no shortage of existing product.

 

[Selaya]

its only a very small reduction from 21gbps to 20 ...
ofc it always feels bad when companies do that but realistically you'd need a microscope to spot the performance delta

 

[sLowEnd]

x70 and x80 soiled forever with the weirdest stack ever. They only needed one generation to do it. Well done Nvidia.

Ever? Naw, there were way more 8800 variants, and with the introduction of G92 caused some naming-performance inconsistencies.

8800 GS 384MB (Had GDDR3 and hobbled DDR2 variants)
8800 GTS 320MB
8800 GTS 640MB
8800 GTS 640MB Core 112
8800 GTS 512MB
8800 GT 256MB (Slower memory clock than standard 512MB version)
8800 GT 512MB (Faster than G80-based GTS due to newer, more capable G92 core)
8800 GTX 768MB
8800 Ultra 768MB
...and I'm pretty sure I still missed a few

 

[lexluthermiester]

This could be a good lower cost model.

This can be fair if done in a consumer respectful way.

Agreed. As long as the price is right and they DON'T gimp it, it'll be a great offering.

EDIT: Left out a key word.. Sleep deprivation is a thing..

 

[basco]

wow i am shocked to learn that there was a 8800 GTS 640MB Core 112 - never knew about that.
was that clear on the package when ya bought one or ?

from hardware canucks:
All in all, there is nothing to differentiate this revised 8800GTS card from any other GTS card produced in the last few months.

 

[Knight47]

They could do the same treatment as with 2060 12GB, but I guess not.

 

[phints]

With the slower VRAM this would make a nice $400 late gen release. But it's Nvidia, so it'll likely end up costing $500 making it pointless when a Super is $600.

 

[Beginner Macro Device]

One simple trick to make NV stop doing that:

Spoiler: Ta-da!

Everyone stops buying their stuff, unless heavily discounted

Regarding the name, RTX 4065 wouldn't be the worst idea for that optical stroke. Or, 4069. No need to higher the clocks, either, they'd better off subbing all 4060 Tis with this much more reasonable device.

 

[Dr. Dro]

If it was 256bit and 16GB Vram. I would have considered it. But let's screw the customer even more? Nah, Nvidia can go get screwed!

The characteristics you want are only present in a product 3 tiers above: the 4070 Ti Super.

wow i am shocked to learn that there was a 8800 GTS 640MB Core 112 - never knew about that.
was that clear on the package when ya bought one or ?

from hardware canucks:
All in all, there is nothing to differentiate this revised 8800GTS card from any other GTS card produced in the last few months.

What's the point in digging up a 18 year old midrange graphics card other than stirring the pot? I genuinely wonder. I'll just name a few: GTX 260 with 216 cores, 560 Ti with 448 cores (this one even used a GF110 chip), GTS 450 and 460 had GF116 subvariants, GTX 1650 has TU116 and TU117 subvariants, RTX 3060 has several subvariants built on its original specification, subvariants with GA103, GA104, 3060 Ti has a GDDR6X subvariant, point is: it is very common for the hardware to undergo small revisions in its configuration targeting roughly the same performance due to extreme demand.

This here is when you get the picture, Nvidia will sell 10+ GeForces for every Radeon AMD manages to shift, and this gap widens in the high-volume segment. The demand is so high they sell every unit they can make, revise the hardware to improve availability and still sell every unit they can make afterwards.

 

[basco]

i know about most older gpu´s but did not know about this 8800 gts 112 vesrion.

it was just out of curiosity and not to offend you Mr. Dr Dr. Dr.. Dro
sorry for even posting and asking questions

 

[Dr. Dro]

i know about most older gpu´s but did not know about this 8800 gts 112 vesrion.

it was just out of curiosity and not to offend you Mr. Dr Dr. Dr.. Dro
sorry for even posting and asking questions

Dang I guess I did come across as rude

Apologies

 

[basco]

i was just driving with my wife(and she drove) one hour trough the hot city and saw your comment and went berserk.
but all is good we all do it sometimes and it was offtopic from my side so sorry from me
love ya all at tpu