Home/All charts/Laptop GPU Efficiency Curves
3DMark Time Spy · Stock Settings · 37 GPUs

Geekerwan Laptop GPU Efficiency Curves (3DMark Time Spy)

Geekerwan laptop GPU efficiency curves cover 37 mobile discrete GPUs tested on 3DMark Time Spy at official stock settings. At 90W, the top three GPUs are RTX 5090, RTX 5080, RTX 4090; at full power, the highest peak score belongs to RTX 5090.

Updated 2026-11-09View source data ↗
PEAK · 3DMark Time Spy25234RTX 5090 @ 175 W

Geekerwan Laptop GPU Efficiency Curves

3DMark Time Spy37 curves · axis 0–208W
Performance (points)GPU power (W)0500010000150002000025000025507510012515017520090W
How to read this chart
Why the line sits at 90W90W sits right in the middle of mainstream gaming-laptop power limits and covers the largest number of tested GPUs (22 of 37). The equal-power table below compares scores along this line.
Every curve uses measured pointsUnlike fitted mobile SoC curves, laptop GPU curves plot actual 3DMark Time Spy runs at each wattage step. Hover any active curve to read its exact power (W) and score.
Stock settings onlyAll GPUs were tested at official default settings with no overclocking or undervolting, isolating real architectural efficiency across wattages.
Under 4 points shown as dotsMatching the upstream rule, GPUs with fewer than 4 tested power steps are drawn as discrete dots without connecting lines. Models whose tested range does not cover 90W show “—” in the equal-power table but still rank in the peak table.
Per wattPeak score ÷ peak power measures efficiency at full TGP. Click the headers in the peak ranking to sort by score or per-watt score.

The two tables below are derived from the upstream dataset (equal-power scores are interpolated along the upstream curves) and provided for reference only. If you believe something is off, email us via the address on the About page.

Equal-power performance (90W)

3DMark Time Spy22 of 37 ranked90W GPU power · higher is better
#GPU90W
01RTX 5090NVIDIA Blackwell18349.00
02RTX 5080NVIDIA Blackwell16528.00
03RTX 4090NVIDIA Ada Lovelace15629.00
04RTX 5070 TiNVIDIA Blackwell14757.00
05RTX 4080NVIDIA Ada Lovelace13830.00
06RTX 5070NVIDIA Blackwell13810.00
07RTX 5060NVIDIA Blackwell11810.00
08RTX 4070 NewNVIDIA Ada Lovelace11715.00
09RTX 4070NVIDIA Ada Lovelace11329.00
10RTX 4060NVIDIA Ada Lovelace10212.00
11RTX 5050NVIDIA Blackwell10157.00
12RTX 3080NVIDIA Ampere9743.00
13RTX 3070 TiNVIDIA Ampere9378.00
14RTX 3080 TiNVIDIA Ampere9371.00
15RTX 3070NVIDIA Ampere9240.00
16RTX 4050NVIDIA Ada Lovelace8664.00
17RTX 2080 Max-QNVIDIA Turing8123.00
18RTX 3060NVIDIA Ampere7822.00
19RTX 2070 SuperNVIDIA Turing7541.00
20RTX 2070NVIDIA Turing6762.00
21RTX 2060NVIDIA Turing6167.00
22RTX 3050NVIDIA Ampere5307.00
—RTX 3050 TiNVIDIA Ampere—
—RTX 2050NVIDIA Ampere—
—RTX 2080 SuperNVIDIA Turing—
—RTX 2080NVIDIA Turing—
—GTX 1650NVIDIA Turing—
—MX550NVIDIA Turing—
—GTX 1080NVIDIA Pascal—
—GTX 1070NVIDIA Pascal—
—GTX 1060NVIDIA Pascal—
—RX 6850M XTAMD RDNA 2—
—RX 6800MAMD RDNA 2—
—RX 6800SAMD RDNA 2—
—RX 6650MAMD RDNA 2—
—RX 6600MAMD RDNA 2—
—RX 6600SAMD RDNA 2—

Full-power peak ranking

3DMark Time Spy37 GPUsDefault: peak score · headers switch to per wattPeak ↓
#GPUPeak power
01RTX 5090NVIDIA Blackwell25234.00175.00 W144.19
02RTX 5080NVIDIA Blackwell22883.00175.00 W130.76
03RTX 4090NVIDIA Ada Lovelace22865.00175.00 W130.66
04RTX 4080NVIDIA Ada Lovelace19185.00175.00 W109.63
05RTX 5070 TiNVIDIA Blackwell18038.00140.00 W128.84
06RTX 5070NVIDIA Blackwell14698.00115.00 W127.81
07RTX 3080 TiNVIDIA Ampere13541.00175.00 W77.38
08RTX 4070 NewNVIDIA Ada Lovelace13064.00140.00 W93.31
09RTX 5060NVIDIA Blackwell12488.00115.00 W108.59
10RX 6850M XTAMD RDNA 212429.00175.00 W71.02
11RTX 3080NVIDIA Ampere12202.00155.00 W78.72
12RTX 4070NVIDIA Ada Lovelace12135.00110.00 W110.32
13RTX 3070 TiNVIDIA Ampere11855.00150.00 W79.03
14RX 6800MAMD RDNA 211595.00174.00 W66.64
15RTX 3070NVIDIA Ampere11366.00150.00 W75.77
16RTX 2080 SuperNVIDIA Turing11226.00200.00 W56.13
17RTX 2080NVIDIA Turing11047.00200.00 W55.23
18RTX 4060NVIDIA Ada Lovelace10533.00110.00 W95.75
19RTX 5050NVIDIA Blackwell10225.00110.00 W92.95
20RX 6800SAMD RDNA 29309.00100.00 W93.09
21RTX 3060NVIDIA Ampere9037.00140.00 W64.55
22RX 6650MAMD RDNA 28790.00120.00 W73.25
23RTX 4050NVIDIA Ada Lovelace8758.0095.00 W92.19
24RTX 2080 Max-QNVIDIA Turing8711.00100.00 W87.11
25RX 6600MAMD RDNA 28133.00120.00 W67.78
26RTX 2070 SuperNVIDIA Turing8033.00115.00 W69.85
27RX 6600SAMD RDNA 27765.0080.00 W97.06
28RTX 2070NVIDIA Turing7454.00115.00 W64.82
29GTX 1080NVIDIA Pascal7288.00200.00 W36.44
30RTX 2060NVIDIA Turing6733.00115.00 W58.55
31RTX 3050 TiNVIDIA Ampere6003.0095.00 W63.19
32GTX 1070NVIDIA Pascal5458.00115.00 W47.46
33RTX 3050NVIDIA Ampere5381.0095.00 W56.64
34RTX 2050NVIDIA Ampere3993.0060.00 W66.55
35GTX 1060NVIDIA Pascal3722.0078.00 W47.72
36GTX 1650NVIDIA Turing3475.0050.00 W69.50
37MX550NVIDIA Turing2893.0050.00 W57.86

How to read laptop GPU efficiency curves

What the two axes mean

The x-axis is laptop GPU power (W) and the y-axis is the 3DMark Time Spy graphics score. All GPUs are tested at official stock settings with no overclocking or undervolting; curves further toward the upper left deliver more performance per watt.

Match the wattage to your laptop class

For thin-and-light creator laptops, focus on 40W–80W; for mainstream gaming laptops, check 80W–120W; for full-size flagships, look at 140W–175W. Newer architectures often match or beat older flagship GPUs while drawing far less power.

Equal power vs full TGP

At full power, generational leaps are just as clear: RTX 5090 reaches 25234 points at 175W, while RTX 5080 scores 22883 at 175W. Match a laptop’s rated GPU TGP against these curves to estimate its real gaming tier.

Glossary

GPU power (TGP / W)
The power budget allocated to a laptop GPU. The same GPU model can be capped at 60W, 100W, or 140W depending on the chassis, creating major performance differences.
3DMark Time Spy
UL’s DirectX 12 graphics benchmark widely used to compare rasterized gaming performance across laptop and desktop GPUs.
Equal-power score (90W)
Interpolated Time Spy graphics score at 90W, removing chassis power-limit differences to compare pure GPU efficiency.
Per watt
Peak score divided by peak power, showing how much performance each watt delivers at the GPU’s maximum tested TGP.

Laptop GPU efficiency FAQ

Why do two laptops with the same GPU model perform so differently?

+

Laptop makers configure different GPU power limits (TGP) based on cooling. These curves show exactly how much performance a GPU delivers at 60W, 90W, or 140W — and where scaling flattens out.

Why do some curves flatten out past 100W–115W?

+

Several mainstream GPUs hit their voltage or clock ceiling around 100W–115W, meaning extra wattage yields almost no extra score. The bend in the curve marks the sweet spot.

Why do some GPUs show “—” in the 90W table?

+

Entry-level thin-and-light GPUs top out below 90W, while some flagship models start testing above 90W. Those models show “—” at 90W but still appear in the full-power peak ranking.

Were any of these GPUs overclocked or undervolted?

+

No. Every laptop GPU was tested at official stock settings with external variables minimized as much as possible.

More Geekerwan charts

Same chips, a different lens.