NVIDIA GeForce 920M vs NVIDIA GeForce RTX 3060 Ti GDDR6X Comparison

NVIDIA
GEFORCE

NVIDIA GeForce 920M

CORE STATE GK208B
VRAM 2 GB
CLOCK SPEED 954 MHz
TDP 33 W
BUS WIDTH 64 bit
ARCHITECTURE Kepler 2.0
nm
PROCESS 28 nm
LAUNCH DATE 2015
VS
NVIDIA
GEFORCE

GeForce RTX 3060 Ti GDDR6X

CORE STATE GA104
VRAM 8 GB
CLOCK SPEED 1665 MHz
TDP 225 W
BUS WIDTH 256 bit
ARCHITECTURE Ampere
nm
PROCESS 8 nm
LAUNCH DATE 2022

PERFORMANCE BENCHMARKS

geekbench_opencl
3,725
N/A
geekbench_vulkan
2,849
N/A
3dmark_3dmark_steel_nomad_dx12
N/A
2,795

Analysis: NVIDIA GeForce 920M vs NVIDIA GeForce RTX 3060 Ti GDDR6X

The NVIDIA GeForce 920M and the NVIDIA GeForce RTX 3060 Ti GDDR6X represent two distant points in NVIDIA’s mobile and desktop GPU timelines. The 920M is a 2015-era entry-level notebook part built on the Kepler architecture, while the RTX 3060 Ti GDDR6X is a 2022 Ampere-based desktop card. The data available for these two GPUs is sparse and uneven—the 920M has two Geekbench scores, while the RTX 3060 Ti GDDR6X has a single 3DMark result. This makes direct comparison challenging, but the specification sheets reveal a chasm in nearly every measurable category, from transistor count to memory bandwidth. The following analysis relies exclusively on the provided facts to interpret what these numbers mean for real-world usage.

FAQ

Q: How do the average benchmark scores of the two GPUs compare?

A: The NVIDIA GeForce 920M has an average benchmark score of 3,287, while the NVIDIA GeForce RTX 3060 Ti GDDR6X has an average benchmark score of 2,795. The 920M’s score is actually 17.6% higher than the RTX 3060 Ti’s score, but this is misleading because the tests are entirely different: the 920M runs Geekbench OpenCL and Vulkan, while the RTX 3060 Ti runs 3DMark Steel Nomad DX12.

Q: Which GPU has a higher transistor density, and what does that imply?

A: The RTX 3060 Ti GDDR6X has a transistor density of 44.4M / mm², compared to 11.7M / mm² for the 920M. This nearly 4x difference reflects the much more modern 8 nm Samsung process node used by the Ampere chip (GA104) versus the 28 nm TSMC node used by the Kepler chip (GK208B).

Q: What are the memory bandwidth figures for each GPU?

A: The 920M has a memory bandwidth of 14.40 GB/s, using 2 GB of DDR3 on a 64-bit bus. The RTX 3060 Ti GDDR6X has a memory bandwidth of 608.3 GB/s, using 8 GB of GDDR6X on a 256-bit bus. That is a 42.2x difference in raw bandwidth.

Q: How do the shading units and texture units differ between the two?

A: The 920M has 384 shading units, 32 TMUs, and 8 ROPs. The RTX 3060 Ti GDDR6X has 4,864 shading units, 152 TMUs, and 80 ROPs. The RTX 3060 Ti has roughly 12.7x more shading units and 4.75x more TMUs.

Q: Do both GPUs support the same DirectX version?

A: No. The 920M supports DirectX 12 (11_0), while the RTX 3060 Ti GDDR6X supports DirectX 12 Ultimate (12_2). The RTX 3060 Ti also adds hardware ray tracing cores (38) and tensor cores (152), which the 920M lacks entirely.

Q: Which GPU has a higher pixel fill rate?

A: The RTX 3060 Ti GDDR6X has a pixel rate of 133.2 GPixel/s, compared to 7.632 GPixel/s for the 920M. This is a 17.5x difference, directly impacting how fast each card can fill the screen with rendered pixels.

Where Each One Wins

The 920M wins in the Geekbench OpenCL and Vulkan tests, scoring 3,725 and 2,849 respectively. Its average benchmark score of 3,287 places it in the 20th percentile of all GPUs, and it sits within 3% of rivals like the Intel HD Graphics 530 (3,332) and the NVIDIA GeForce GT 730M (3,316). This suggests the 920M is competitive in lightweight, compute-oriented workloads that are typical of older or less demanding applications, such as basic media playback, legacy game titles, or OpenCL-accelerated productivity tasks.

The RTX 3060 Ti GDDR6X wins in the 3DMark Steel Nomad DX12 test with a score of 2,795. Its percentile ranking is 18, which is lower than the 920M’s 20th percentile, but the workload is radically different—Steel Nomad is a modern DX12 benchmark that stresses multi-threaded rendering, ray tracing, and high-bandwidth memory. The RTX 3060 Ti’s nearest rivals include the AMD Radeon 550 (2,801, -0.2%) and the NVIDIA GeForce RTX 4010 (2,893, +3.4%), indicating it trades blows with entry-level desktop cards in this specific test. The 920M has no ray tracing support, so it simply cannot participate in the same class of workloads.

Architecture Differences

The 920M is built on the Kepler 2.0 architecture, using the GK208B chip fabricated on a 28 nm process at TSMC. It packs 1,020 million transistors into an 87 mm² die. The RTX 3060 Ti GDDR6X uses the Ampere architecture, with the GA104 chip built on an 8 nm process at Samsung. It contains 17,400 million transistors on a 392 mm² die. The transistor density jump from 11.7M / mm² to 44.4M / mm² is stark, reflecting over six years of process node advancement.

The 920M has no ray tracing cores and no tensor cores. The RTX 3060 Ti GDDR6X has 38 ray tracing cores and 152 tensor cores, enabling hardware-accelerated ray tracing and AI-based features like DLSS. The 920M supports DirectX 12 (11_0), while the RTX 3060 Ti supports DirectX 12 Ultimate (12_2), which includes features like mesh shaders and variable-rate shading. Vulkan support also differs: the 920M reports Vulkan 1.2.175, while the RTX 3060 Ti reports Vulkan 1.4. Both support OpenGL 4.6.

Specification Differences

The two GPUs differ in nearly every specification field. The 920M has a base and boost clock of 954 MHz, while the RTX 3060 Ti GDDR6X has a base clock of 1,410 MHz and a boost clock of 1,665 MHz. Memory is a major divergence: the 920M uses 2 GB of DDR3 at 1,800 Mbps effective, while the RTX 3060 Ti uses 8 GB of GDDR6X at 19 Gbps effective. The bus width jumps from 64-bit to 256-bit, and bandwidth from 14.40 GB/s to 608.3 GB/s.

Compute throughput is drastically different. The 920M delivers 732.7 GFLOPS of FP32 performance, whereas the RTX 3060 Ti GDDR6X delivers 16.20 TFLOPS. The RTX 3060 Ti also has FP16 performance of 16.20 TFLOPS (1:1), while the 920M has no listed FP16 figure. Texture rate is 30.53 GTexel/s for the 920M versus 253.1 GTexel/s for the RTX 3060 Ti. Pixel rate is 7.632 GPixel/s versus 133.2 GPixel/s. The TDP tells the story of power: the 920M is rated at 33 W, while the RTX 3060 Ti is rated at 225 W, requiring a 550 W suggested PSU and a 12-pin power connector. The 920M is an IGP with no power connectors, while the RTX 3060 Ti is a dual-slot card measuring 242 mm in length and 112 mm in height.

Head-to-Head Benchmarks

There are no direct head-to-head benchmark results shared between the two GPUs, and the available tests are not comparable. The 920M’s Geekbench OpenCL score of 3,725 and Vulkan score of 2,849 are compute-focused, measuring general-purpose GPU acceleration. Its average score of 3,287 places it just 0.9% below the GeForce GT 730M (3,316) and 1.4% below the Intel HD Graphics 530 (3,332). This shows the 920M is firmly in the entry-level mobile segment, competing with integrated graphics from the same era.

The RTX 3060 Ti GDDR6X’s single 3DMark Steel Nomad DX12 score of 2,795 is a modern gaming benchmark. Its nearest rivals include the AMD Radeon 550 (2,801, -0.2%) and the NVIDIA GeForce 820M (2,784, +0.4%), which are also low-end cards. The RTX 3060 Ti GDDR6X is 1.6% ahead of the Intel UHD Graphics 610 (2,752) but 3.4% behind the NVIDIA GeForce RTX 4010 (2,893). This places the RTX 3060 Ti GDDR6X in an odd position: despite its massive hardware specifications, its 3DMark result is not far from entry-level parts, likely because the Steel Nomad test is not representative of its full capabilities.

The only clear wins are in the specification sheet. The RTX 3060 Ti GDDR6X is 22.1x faster in FP32 compute (16.20 TFLOPS vs 732.7 GFLOPS), has 42.2x more memory bandwidth (608.3 GB/s vs 14.40 GB/s), and 17.5x higher pixel rate (133.2 GPixel/s vs 7.632 GPixel/s). It also has 38 ray tracing cores and 152 tensor cores, which the 920M lacks entirely. In contrast, the 920M has a lower TDP (33 W vs 225 W), making it suitable for fanless or low-power designs, and it supports PCIe 3.0 x8, while the RTX 3060 Ti uses PCIe 4.0 x16.

The Verdict

The data paints an unequivocal picture: the NVIDIA GeForce RTX 3060 Ti GDDR6X is vastly more capable than the NVIDIA GeForce 920M in raw performance metrics. Its 8 GB of GDDR6X memory, 608.3 GB/s bandwidth, and 16.20 TFLOPS of FP32 compute put it in a different performance class entirely. The 920M’s 2 GB of DDR3 and 732.7 GFLOPS are suited only to the lightest workloads, and its 20th percentile ranking among all GPUs confirms it as an entry-level part. The RTX 3060 Ti GDDR6X’s 18th percentile in the 3DMark Steel Nomad test is misleading, as that single test does not capture its ray tracing or tensor core advantages.

Who should pick the 920M? Users constrained to a low-power, integrated form factor—the 33 W TDP and IGP slot width make it a candidate for thin-and-light laptops that never see demanding 3D applications. Its Geekbench scores are respectable for its class, trading blows with the GT 730M and HD Graphics 530, so it can handle basic OpenCL acceleration and older APIs.

Who should pick the RTX 3060 Ti GDDR6X? Anyone running modern DX12 Ultimate titles, ray-traced games, or GPU-accelerated AI workloads. The 38 ray tracing cores and 152 tensor cores are not present in the 920M at all, and the 8 GB of GDDR6X memory is four times larger. The 225 W TDP and dual-slot cooler demand a proper desktop PSU, but the performance delta in pixel rate, texture rate, and memory bandwidth is overwhelming. The benchmark data is sparse, but the specification differences alone justify choosing the RTX 3060 Ti GDDR6X for any task beyond legacy 2D or light compute.

DETAILED SPECIFICATIONS

SPECIFICATION
920M
RTX 3060 Ti GDDR6X
Core Specs
Shading Units
384
4,864 +1166.7%
Shaders
384
4,864 +1166.7%
TMUs
32
152 +375.0%
ROPs
8
80 +900.0%
SM Count
—
38
Clocks
Base Clock
954 MHz
1410 MHz
Boost Clock
954 MHz
1665 MHz
Memory Clock
900 MHz 1800 Mbps effective
1188 MHz 19 Gbps effective
Memory
Memory Size
2 GB
8 GB
VRAM (MB)
2,048
8,192 +300.0%
Memory Type
DDR3
GDDR6X
Memory Bus
64 bit
256 bit
Bandwidth
14.40 GB/s
608.3 GB/s
Cache
L1 Cache
16 KB (per SMX)
128 KB (per SM)
L2 Cache
512 KB
4 MB
Performance
Pixel Rate
7.632 GPixel/s
133.2 GPixel/s
Texture Rate
30.53 GTexel/s
253.1 GTexel/s
FP32 (TFLOPS)
732.7 GFLOPS
16.20 TFLOPS
FP64 (TFLOPS)
30.53 GFLOPS (1:24)
253.1 GFLOPS (1:64)
FP16 (TFLOPS)
—
16.20 TFLOPS (1:1)
AI/RT
RT Cores
—
38
Tensor Cores
—
152
Power
TDP
33 W
225 W
TDP (W)
33
225 +581.8%
Suggested PSU
—
550 W
Power Connectors
None
1x 12-pin
Architecture
Architecture
Kepler 2.0
Ampere
GPU Name
GK208B
GA104
Generation
GeForce 900M
GeForce 30
Process Size
28 nm
8 nm
Transistors
1,020 million
17,400 million
Die Size
87 mm²
392 mm²
Foundry
TSMC
Samsung
Density
11.7M / mm²
44.4M / mm²
API Support
DirectX
12 (11_0)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.2.175
1.4
OpenCL
3.0
3.0
CUDA
3.5
8.6
Shader Model
6.5 (5.1)
6.8
Physical
Slot Width
IGP
Dual-slot
Length
—
242 mm 9.5 inches
Height
—
112 mm 4.4 inches
Outputs
Portable Device Dependent
1x HDMI 2.13x DisplayPort 1.4a
Bus Interface
PCIe 3.0 x8
PCIe 4.0 x16
Other
Production
End-of-life
End-of-life
Predecessor
GeForce 800M
GeForce 20
Successor
GeForce 10 Mobile
GeForce 40
View GeForce 920M Details View GeForce RTX 3060 Ti GDDR6X Details