NVIDIA RTX 3080 Ti and RTX 3070 Ti, characteristics of these GPUs

Within the vary of NVIDIA merchandise based mostly on its RTX 30 structure or often called GeForce Ampere, there have been two gaps to fill as a result of distance in value between the RTX 3070 and the RTX 3080 on the one hand and from the RTX 3080 to the RTX 3090 The new RTX 3070 Ti is positioned within the first gap and the RTX 3080 Ti within the second. Thus finishing the mid and excessive ranges of the NVIDIA RTX 30.

NVIDIA GeForce RTX 3080 Ti

RTX 3080 Ti NVIDIA

Graphic card NVIDIA RTX 3080 Ti
GPU GA102-225
Family RTX 30
Architecture GeForce Ampere
SM / CUs 80
FP32 ALUs by SM / CU 128
Ray Tracing Unit Yes, 1 per SM (80 complete)
Tensor Units Yes, Four per SM
TFLOPS FP32 3. 4
TFLOPS Tensioner (FP16) 273
VRAM sort GDDR6X
VRAM amount 12 GB
VRAM pace 19 Gbps
VRAM bandwidth 912 GB / s
PVPR $ 1199
Base pace 1365 MHz
Boost pace 1665 MHz

The first and strongest of the brand new NVIDIA graphics playing cards is the GeForce RTX 3080 Ti, which is the third mannequin based mostly on the NVIDIA GA102 GPU after the RTX 3080 and RTX 3090. In this case it’s based mostly on the chip GA102-225

Starting with VRAM, the graphics card has a 384-bit GDDR6X bus just like the RTX 3090, however with 12 GB capability, since NVIDIA has determined to not use the x8 mode of and due to this fact has put in half the reminiscence chips than in its older sister. Another change is the considerably slower reminiscence utilization, from 19 Gbps as an alternative of 19.5 Gbps. So it makes use of the identical sort of GDDR6X because the GeForce 3080. However, its consumption is 350 W, the identical as that of the RTX 3090 and 30 W above the 320 W of the RTX 3080.

If we go to the GPU itself, we’re confronted with a 80 SM configuration to a 1365 MHz base pace with a 1665 MH enhancez, which supplies it an influence of 34 TFLOPS in FP32, 67 RT-FLOPS and 274 Tensor TFLOPS. The first measure comes from the calculation capability of the 128 ALUs in FP32 of every of the SMs, the second is a comparative measure created by NVIDIA for its graphics playing cards and the third is the calculation capability of the Tensor Cores.

As for its value, it has a prompt retail value of $ 1,199, which places it nearly double the RTX 3080’s MSRP of $ 699.

NVIDIA RTX 3070 Ti

RTX 3070 Ti

Graphic card NVIDIA RTX 3080 Ti
GPU GA104-400
Family RTX 30
Architecture GeForce Ampere
SM / CUs 48
FP32 ALUs by SM / CU 128
Ray Tracing Unit Yes, 1 per SM (48 complete)
Tensor Units Yes, Four per SM
TFLOPS FP32 22
TFLOPS Tensioner (FP16) 174
VRAM sort GDDR6X
VRAM amount eight GB
VRAM pace 19 Gbps
VRAM bandwidth 609 GB / s
PVPR $ 1199
Base pace 1580 MHz
Boost pace 1770 MHz

The second mannequin that NVIDIA has offered at Computex 2021 is the RTX 3070 Ti, which instantly turns into essentially the most highly effective mid-range graphics card within the RTX 30. Since just like the RTX 3060 Ti and RTX 3070 Ti that is based mostly on a variant of the GA104, particularly the GA104-400.

Compared to the RTX 3070 this has the complete configuration of 48 SM, along with working at a 1580 MHz base clock pace and 1770 MHz Boost, that are increased than the RTX 3070. Biggest shock? NVIDIA has changed the 256-bit GDDR6 reminiscence interface with one of the identical width GDDR6X so the capability of the VRAM remains to be eight GB, however in change it has gone from having 14 Gbps reminiscence to 19 Gbps, rising its bandwidth from 448 GB / s to 609 GB / s. However, this improve in bandwidth comes with a rise within the card’s vitality consumption, which matches from 220 W to 290 W complete.

Regarding the technical specs, we discover a energy of 22 TFLOPS in FP32, 42 RT-TFLOPS and 174 Tensor TFLOPS. The which means of the three efficiency parameters being the identical that we now have mentioned within the RTX 3080 Ti and due to this fact the identical as the remaining of the NVIDIA RTX 30 vary.

The NVIDIA beneficial retail value for the RTX 3070 Ti? 599 {dollars}.

Based on GeForce Ampere structure

GeForce Ampere RTX 30

Like the remaining of the RTX 30s which have appeared available on the market in latest months, each the RTX 3080 Ti and RTX 3070 Ti are based mostly on the GeForce Ampere structure, which depends on three completely different pillars to realize its huge energy:

  • Improved CUDA cores, now with a second set of FP32 ALUs switched with the integer ALUs, permits you to carry out as much as twice the quantity of operations per SM and clock cycle than GeForce Turing, aka RTX 20s.
  • New RT Cores, the Ray Tracing functionality has been improved, to the purpose of doubling the efficiency within the face of the beam-box and beam-triangle intersection.
  • Third Generation Tensor Cores, which permits an acceleration of DLSS algorithms and to have the ability to use NVIDIA Broadcast at an incredible pace and with out affecting the efficiency of the sport.

These three parts immediately make the RTX 30 graphics structure essentially the most highly effective available on the market up to now, making each the RTX 3080 Ti and RTX 3070 Ti graphics with unprecedented energy of their respective value ranges.

To all this we should additionally add help for NVIDIA Reflex, which is designed to scale back gaming latencies as a lot as doable, permitting refresh charges of as much as 360 Hz on G-SYNC displays.

RTX 3070 Ti and RTX 3080 Ti be a part of the 2nd of the RTX 30

NVIDIA Resizable BAR

NVIDIA has utterly renewed its RTX 30 in a second technology that’s designed particularly for the gaming market. The approach to obtain this has been by slicing Ethereum’s mining capability in half, a change that has been carried out by NVIDIA additionally within the RTX 3080 Ti and the RTX 3070 Ti.

Another change that NVIDIA has carried out is the serial help for the usual ResizableBAR, this know-how consists of giving entry to the VRAM to the CPU. Its utility? It permits GPUs to learn info from command lists a lot quicker by not having to entry system RAM to acquire them. Which interprets right into a reduce in body time and due to this fact extra FPS in video games.