-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
NVIDIA 4000 Series
- Thread starter MrClippy
- Start date
Just having any cards available for purchase would be a bonus.
Absolutely
The reason why AIO liquid cooling will be common next gen is cooling itself of a premium product. The whole point of a premium card is that it runs faster and cooler than entry level but because nextbgen cards runs so hot no one will be able to stick a tiny cheap cooler on a RTx4090 and so if someone like ASUS tries to make all its RTX4090's air cooled there will be almost no difference in performance or temperature between the models simply because there is no room for product segmentation, every air cooled 4090 is going to run at 80c doesn't matter if it's entry level or a Strix
That's where water cooling comes in, it's the same only way they'll be able to segment their product stack by offering cards that run faster and cooler than entry level models so yes I absolutely 100% believe there will be lots Of liquid cooled cards next generation. And if somehow I'm completely wrong and all they sell is 600w entry and high end air cooled 4090s then there will be zero reason to buy anything other than the entry level model and you can quote me on this.
I know I'm right as long as the 600w rumours are accurate because the 3090ti is only 500w and we already see that happening, most of the high end 3090ti's are AIO liquid cooled
That's where water cooling comes in, it's the same only way they'll be able to segment their product stack by offering cards that run faster and cooler than entry level models so yes I absolutely 100% believe there will be lots Of liquid cooled cards next generation. And if somehow I'm completely wrong and all they sell is 600w entry and high end air cooled 4090s then there will be zero reason to buy anything other than the entry level model and you can quote me on this.
I know I'm right as long as the 600w rumours are accurate because the 3090ti is only 500w and we already see that happening, most of the high end 3090ti's are AIO liquid cooled
I feel GPU's with a built in AIO would be far more popular were they to make the pump RPM adjustable. All the cards I tried and seen have had "noisy" pumps that run at full speed all the time (or close to it). Having 0db at idle is a feature sought after by many and is obviously achievable on air with fan stop.
Also, many AIO manufactuers include a fan on the GPU even with an external radiator. This smaller fan is usually very loud, as it's such a small diameter and has to spin and high RPM to do anything.
Hope some AIO's, NV and AMD adopt the same design as on the Vega64 Liquid cooled, but with an adjustable pump RPM and design to account for a low flow rate at idle. I owned the Vega64 LC, excellent cooling system and design, would have been perfect with a lower pump speed at idle.
Makes sense really and fingers crossed this happens as I have been looking forward to getting a liquid cooled gpu for a while now!
Associate
Currently £700 for a 3070 Ti, yeah expect the 4000 series to be sold out immediately and cue NVidia saying they had no idea there would be so much demand. To be fair they had to order the silicon chips a while ago, its just a guess. Prices will be around £500-600 for a 4070 and then £600-700 when the Ti comes out.
Kopeite says he has the AD102's SM architecture block diagrams and details:
The NVIDIA Ada Lovelace AD102 GPU will feature up to 12 GPC (Graphics Processing Clusters). This is an increase of 70% versus GA102 which features only 7 GPCs. Each GPC will consist of 6 TPCs and 2 SMs which is the same configuration as the existing chip. Each SM (Streaming Multiprocessor) will house four sub-cores which is also the same as the GA102 GPU. What's changed is the FP32 & the INT32 core configuration. Each sub-core will include 128 FP32 units but combined FP32+INT32 units will go up to 192. This is because the FP32 units don't share the same sub-core as the IN32 units. The 128 FP32 cores are separate from the 64 INT32 cores.
So in total, each sub-core will consist of 128 FP32 plus 64 INT32 units for a total of 192 units. Each SM will have a total of 512 FP32 units plus 256 INT32 units for a total of 768 units. And since there are a total of 24 SM units (2 per GPC), we are looking at 12,288 FP32 Units and 6,144 INT32 units for a total of 18,432 cores. Each SM will also include two Wrap Schedules (32 thread/CLK) for 64 wraps per SM. This is a 50% increase on the cores (FP32+INT32) and a 33% increase in Wraps/Threads vs the GA102 GPU.
Moving over to the cache, this is another segment where NVIDIA has given a big boost over the existing Ampere GPUs. The Ada Lovelace GPUs will pack 192 KB of L1 cache per SM, an increase of 50% over Ampere. That's a total of 4.5 MB of L1 cache on the top AD102 GPU. The L2 cache will be increased to 96 MB. This is a 16x increase over the Ampere GPU that hosts just 6 MB of L2 cache. The cache will be shared across the GPU.
Finally, we have the ROPs which are also increased to 32 per GPC, an increase of 2x over Ampere. You are looking at 384 ROPs on the AD102.
wccftech.com
- 70% more GPC units
- 50% More FP32 Cuda Cores
- 50% More L1 Cache
- 16x More L2 Cache
- 2x more ROP cores per GPC
- 4th Gen Tensor & 3rd Gen RT Cores
The NVIDIA Ada Lovelace AD102 GPU will feature up to 12 GPC (Graphics Processing Clusters). This is an increase of 70% versus GA102 which features only 7 GPCs. Each GPC will consist of 6 TPCs and 2 SMs which is the same configuration as the existing chip. Each SM (Streaming Multiprocessor) will house four sub-cores which is also the same as the GA102 GPU. What's changed is the FP32 & the INT32 core configuration. Each sub-core will include 128 FP32 units but combined FP32+INT32 units will go up to 192. This is because the FP32 units don't share the same sub-core as the IN32 units. The 128 FP32 cores are separate from the 64 INT32 cores.
So in total, each sub-core will consist of 128 FP32 plus 64 INT32 units for a total of 192 units. Each SM will have a total of 512 FP32 units plus 256 INT32 units for a total of 768 units. And since there are a total of 24 SM units (2 per GPC), we are looking at 12,288 FP32 Units and 6,144 INT32 units for a total of 18,432 cores. Each SM will also include two Wrap Schedules (32 thread/CLK) for 64 wraps per SM. This is a 50% increase on the cores (FP32+INT32) and a 33% increase in Wraps/Threads vs the GA102 GPU.
Moving over to the cache, this is another segment where NVIDIA has given a big boost over the existing Ampere GPUs. The Ada Lovelace GPUs will pack 192 KB of L1 cache per SM, an increase of 50% over Ampere. That's a total of 4.5 MB of L1 cache on the top AD102 GPU. The L2 cache will be increased to 96 MB. This is a 16x increase over the Ampere GPU that hosts just 6 MB of L2 cache. The cache will be shared across the GPU.
Finally, we have the ROPs which are also increased to 32 per GPC, an increase of 2x over Ampere. You are looking at 384 ROPs on the AD102.
NVIDIA Ada Lovelace 'GeForce RTX 40' Gaming GPU Detailed: Double The ROPs, Huge L2 Cache & 50% More FP32 Units Than Ampere, 4th Gen Tensor & 3rd Gen RT Cores
Details regarding the NVIDIA Ada Lovelace Gaming GPU which will power the GeForce RTX 40 series graphics cards have been revealed.
wccftech.com
Last edited:
Those sound like monsters. You sure you are going to be able to fit three in?Kopeite says he has the AD102's SM architecture block diagrams and details:
- 70% more GPC units
- 50% More FP32 Cuda Cores
- 50% More L1 Cache
- 16x More L2 Cache
- 2x more ROP cores per GPC
- 4th Gen Tensor & 3rd Gen RT Cores
The NVIDIA Ada Lovelace AD102 GPU will feature up to 12 GPC (Graphics Processing Clusters). This is an increase of 70% versus GA102 which features only 7 GPCs. Each GPC will consist of 6 TPCs and 2 SMs which is the same configuration as the existing chip. Each SM (Streaming Multiprocessor) will house four sub-cores which is also the same as the GA102 GPU. What's changed is the FP32 & the INT32 core configuration. Each sub-core will include 128 FP32 units but combined FP32+INT32 units will go up to 192. This is because the FP32 units don't share the same sub-core as the IN32 units. The 128 FP32 cores are separate from the 64 INT32 cores.
So in total, each sub-core will consist of 128 FP32 plus 64 INT32 units for a total of 192 units. Each SM will have a total of 512 FP32 units plus 256 INT32 units for a total of 768 units. And since there are a total of 24 SM units (2 per GPC), we are looking at 12,288 FP32 Units and 6,144 INT32 units for a total of 18,432 cores. Each SM will also include two Wrap Schedules (32 thread/CLK) for 64 wraps per SM. This is a 50% increase on the cores (FP32+INT32) and a 33% increase in Wraps/Threads vs the GA102 GPU.
Moving over to the cache, this is another segment where NVIDIA has given a big boost over the existing Ampere GPUs. The Ada Lovelace GPUs will pack 192 KB of L1 cache per SM, an increase of 50% over Ampere. That's a total of 4.5 MB of L1 cache on the top AD102 GPU. The L2 cache will be increased to 96 MB. This is a 16x increase over the Ampere GPU that hosts just 6 MB of L2 cache. The cache will be shared across the GPU.
Finally, we have the ROPs which are also increased to 32 per GPC, an increase of 2x over Ampere. You are looking at 384 ROPs on the AD102.
NVIDIA Ada Lovelace 'GeForce RTX 40' Gaming GPU Detailed: Double The ROPs, Huge L2 Cache & 50% More FP32 Units Than Ampere, 4th Gen Tensor & 3rd Gen RT Cores
Details regarding the NVIDIA Ada Lovelace Gaming GPU which will power the GeForce RTX 40 series graphics cards have been revealed.
i just hope the hype surrounding the 4000 series turns out to be true. i'm not buying unless they give atleast 25% fps increase over my 3080ti.
how are you planning to secure a 4000 series gpu? i'm signed up to Telegram alerts that ping me when 3000 series FE cards are available, it seems **** is the only place that sells them (maybe more retailers do, but its only **** i get told by Telegram). I bought my 3080ti FE this way, hope the 4000 series will work the same way (i'll get pinged).
how are you planning to secure a 4000 series gpu? i'm signed up to Telegram alerts that ping me when 3000 series FE cards are available, it seems **** is the only place that sells them (maybe more retailers do, but its only **** i get told by Telegram). I bought my 3080ti FE this way, hope the 4000 series will work the same way (i'll get pinged).
Last edited by a moderator:
Associate
Nothing suggests so far that they are going to use less leccy so heat isn’t likely to be less either, so dumping that heat is going to likely need as much space as what goes before like you suggest.
My guess is they’ll be whoppers.
My guess is they’ll be whoppers.
Me? I just ask the store where I shop to hold one for me, skips all the queues and don't have to refresh web pages all day and night trying to press buy like everyone else - the perks of knowing the owner
Last edited by a moderator:
FE is cheaper though ?Me? I just ask the store where I shop to hold one for me, skips all the queues and don't have to refresh web pages all day and night trying to press buy like everyone else - the perks of knowing the owner
At this moment they are the only ones that handle founders edition cards
Personally didn't take long getting FE card they did drops more often at that time seems they have really reduced the drops now
Ideally I wanted AIB but just was way overpriced over FE , this time will try for AIB on release but depends how much difference in price might go FE again
Won't rule out AMD but they don't do UK drops maybe that will change for next release ?
Last edited by a moderator:
With the price of cards now aio's should be refillable so they're not unusable after several years. iirc it only took fury x cards around 4 years to start showing lower liquid levels making them dodgy to use. Not like adding a port to the rad to at least give that option would cost much either.
Deleted member 76686
D
Deleted member 76686
I'd be really happy if they ever manage to make a mid-high level graphics card not cost more than a whole console.
With Ampere Nvidia did not sell FE cards in my country so FE price doesn't mean anything for me, we can only buy AIB cards