Nvidia GeForce RTX 5080 laptop GPU impressively outperforms RTX 4080 in synthetic benchmark

[Redaktion]

A new GeForce RTX 5080 laptop 3D Mark Time Spy score has emerged online. In the graphics segment, it outperforms its Ada Lovelace counterpart by over 20%.

https://www.notebookcheck.net/Nvidia-GeForce-RTX-5080-laptop-GPU-impressively-outperforms-RTX-4080-in-synthetic-benchmark.957558.0.html

[Cannolicchio]

20% is not so impressive

[KulkieT]

20% is impressive as long it same power draw

[Mr Majestyk]

4080 mobile gpu is a 4070 Ti (not super) desktop with power limits and fewer cores effectively, while 5080 laptop has more cuda cores and is on 256bit bus vs 192 bit bus, more memory and nearly double the bandwidth. Also 5080 has a 150W TDP limit vs 4080's 110W TDP limit.

So we need to know what TDP's were applied to each of these gpu's but 20% seems low unless it was power starved.

[Calipha]

4080 mobile gpu is a 4070 Ti (not super) desktop with power limits and fewer cores effectively, while 5080 laptop has more cuda cores and is on 256bit bus vs 192 bit bus, more memory and nearly double the bandwidth. Also 5080 has a 150W TDP limit vs 4080's 110W TDP limit.

So we need to know what TDP's were applied to each of these gpu's but 20% seems low unless it was power starved.

The 4080 Laptop GPU being akin to that of the Desktop 4070 Ti isn't too relevant of a factor, as the 5080 Laptop GPU actually has less CUDA cores, SM's, TMU's, and ROP's than the Desktop 5070 Ti, and both the 4080 Laptop and 5080 Laptop GPU have max TGP's of 150W+25W. The 24% increase in Timespy score is based upon the accumulated average of 4080 Laptop GPU's tested in various laptops by Notebookcheck—the average results in a lower than typical score due to several power-constrained 4080 Laptop GPU's being tested. At the same time, the overall and CPU scores in this leaked benchmark lead me to believe that this is a power-constrained laptop, as the scores somewhat align with the Overall/CPU Scores of the Zephyrus G16.

If if is indeed the 2025 variant of the Zephyrus G16, than that would mean that the 5080 Laptop GPU being tested is running in a TGP configuration of 105W+15W. The 4090 Laptop GPU at similar TGP constraints @ 105W+10W receives a TimeSpy Graphics score typically between 16.9K and 17.1K in the same test—resulting in the 5080M having just under a 30% lead, which is quite respectable. It's also possible that it might be even further power-constrained by it's fan profile, as the some stock Ultra 9 185H/4090 Laptop configurations seem to result in higher Overall and CPU scores.

[Zack]

I bought rtx 5080 from bestbuy and nivazon, very satisfied.

[Darkentbg]

4080 mobile gpu is a 4070 Ti (not super) desktop with power limits and fewer cores effectively, while 5080 laptop has more cuda cores and is on 256bit bus vs 192 bit bus, more memory and nearly double the bandwidth. Also 5080 has a 150W TDP limit vs 4080's 110W TDP limit.

So we need to know what TDP's were applied to each of these gpu's but 20% seems low unless it was power starved.

Well yes, but actually no...
My 4080 gets 175W and it doesnt throttle!
Can you guess my score?

[Darkentbg]

A new GeForce RTX 5080 laptop 3D Mark Time Spy score has emerged online. In the graphics segment, it outperforms its Ada Lovelace counterpart by over 20%.

My score is 19 446 with a 4080 - no manual OC, just running the benchmark
I could even push it higher