Nvidia GeForce RTX Mobile in the test

GeForce RTX 2070 Max-Q im Test: Turing für schlanke Notebooks macht keinen Klassensprung

GeForce RTX with Turing GPU is now also available in notebooks. In the test, the GeForce RTX 2070 Max-Q has to assert itself against the predecessors GeForce GTX 1070 Max-Q and GeForce GTX 1080. The performance jump known from the desktop does not occur.

Turing Mobile with and without Max-Q

Turing Mobile is now available in various versions in notebooks from numerous manufacturers. The GeForce RTX 2080 and RTX 2070 are each also available in a Max-Q variant that is specially designed for efficiency, although such a branch of the GeForce RTX 2060 is not currently planned.

All three GPUs have the same configuration as the desktop derivatives. Storage volume and speed are also the same. The clock speeds of the GPU make the difference. According to Nvidia, the turbo of the conventional mobile variants is at least 210 to 400 MHz below that of the desktop version, the Max-Q versions are again up to 500 MHz below.

Gigabyte Aero 15 vs. Razer Blade 15

BitcoinMinersHashrate had a GeForce RTX 2070 Max-Q available for a first test of the new generation. The GPU is used in the Gigabyte Aero 15-X9 with Intel Core i7-8750H (6K / 12T, 45 watts), 32 GB DDR4-2666, and 1 TB NVMe SSD. In this case, the 15.6-inch 60 Hz UHD display from AUO was installed, alternatively there is a Full HD display from LG with 144 Hz. Both are calibrated ex works.

The Gigabyte Aero 15-X9 (Image: Gigabyte)

Even with GeForce RTX, Gigabyte's Aero series lives up to its reputation: With 356 × 250 × 19 mm, the notebook is very compact in relation to the hardware used and, at 2.0 kilograms, it is also comparatively light. Also record-breaking again: the almost 95 watt-hour battery.

RTX 2070 Max-Q vs. GTX 1070 Max-Q

For the direct comparison, the current Razer Blade 15 (test) was used, which has the same CPU and the direct GPU predecessor, the GeForce GTX 1070 Max-Q. The notebook is even a bit thinner than the Aero X9.

Both slim notebooks are therefore ideal for a direct comparison of the generations. However, it would be perfect if the same model were only used with a different graphics card. However, Razer has not yet introduced a new Blade 15 and Gigabyte could not provide the previous year's model from the Aero series.

Comparison of other systems

The following test contains another notebook, the Gigabyte Aorus X7 DT. It relies on the Core i7-8850H and a mobile GeForce GTX 1080 (not Max-Q) and thus represents the next higher performance class of the Pascal era. However, the CPU, which is only 200 MHz faster on paper, also plays a different role in the X7 League than the Core i7-8750H in the other two models because it can consume significantly more power under load: In Kingdom Come: Deliverance (Full HD, details "Ultra-high"), after one hour of load, the clock is 38 watts package. Power at 3.3 GHz, for example, while the other two notebooks clock their CPUs at around 25 watts a good 800 MHz lower. Some of the benchmarks below will be affected.

This also applies to the measured values ​​of the desktop graphics cards: In this case, the test system uses a Core i7-8700K overclocked to 4.7 GHz on all six cores and thus a CPU with a significantly higher clock speed – the difference to the two lies in benchmarks Core i7-8750H partly over 2 GHz. Nevertheless, these values ​​are included to give a better overview of what is possible in desktop and notebook.

test results

Temperature- and consumption-dependent turbo modes make benchmarks of current graphics cards much more complicated even in desktop PCs. In notebooks, these mechanisms and the fact that the processor does not usually deliver the same clock rates over the long term, have an even greater impact. All notebooks, if available, were operated with their standard profile and not with a manually selectable gaming mode. The AI ​​functionality of the Aero 15-X9 was deactivated.

"Cold and warm" test results

In the following benchmarks, BitcoinMinersHashrate compared all three notebooks with measurements from a first run of the benchmarks as well as another one after warming up the GPU and CPU. The Gigabyte Aero 15-X9 with GeForce RTX 2070 Max-Q was most clearly affected by performance losses due to temperature influences. The following diagram shows the course, temperature and clock under constant load in the game Kingdom Come: Deliverance (Full HD, details "Ultra high") of the three notebooks in the test.

Razer Blade 15 and Aorus X7 DT are hardly affected by rising temperatures. The slim notebook from Razer hangs almost permanently in the power target, the larger model from Gigabyte allows almost the highest clock. The Aero 15-X9, on the other hand, has to drop a lot, with around 1,000 MHz, the clock of 1,185 MHz, which Nvidia calls a typical boost, is even undercut.

The reason is the temperature of the GPU, which quickly reaches the critical 86 ° C. The chip is then clocked down, the consumption of the system (without display) drops from 154 to 120 watts. The other notebooks, on the other hand, constantly need around 140 (Razer) or 260 watts (Aorus).

Read This Now:   Here is the GeForce RTX 2060 12 GB, a Russian modder created it before NVIDIA

However, the performance loss in most benchmarks is not as strong as the clock loss suggests. Firstly, the first run is usually affected by falling clock rates and secondly, the influence of the notebook CPU is always the same.

Full HD and Ultra HD performance

The performance leap from GeForce GTX 1070 Max-Q to GeForce RTX 2070 Max-Q in comparison of the two notebooks from Gigabyte and Razer is smaller than known from the desktop PC. On average over the games F1 2018, Kingdom Come: Deliverance and Shadow of the Tomb Raider, Turing is only 12 percent better than Pascal in Full HD with maximum details. If both laptops are warm and the clock speeds drop, especially in the Gigabyte model, there is still a 7 percent advantage.

To edit

Performance rating Full HD

    • GeForce GTX 1080 Mobile (cold)

    • GeForce GTX 1080 Mobile (warm)

    • GeForce RTX 2070 Max-Q (cold)

    • GeForce RTX 2070 Max-Q (warm)

    • GeForce GTX 1070 Max-Q (cold)

    • GeForce GTX 1070 Max-Q (warm)

Far Cry 5 does not fit in this picture (not in the performance rating), because the GeForce GTX 1070 Max-Q is even in the front here. However, this is due to the CPU, which clocks lower in the Aero 15-X9 than in the Razer Blade 15, which doesn't suit the game at all.

In Ultra HD, Turing can stand out from Pascal with 16 percent on average over the three games mentioned, as long as the clock in the aero has not yet fallen. Then again there is only an 8 percent lead.

In this case, the significantly higher load on the GPU reduces the influence of the CPU and the Aorus X7 DT in particular is much stronger. With the lower CPU speed, the GeForce RTX 2070 Max-Q in the Aero 15-X9 in Far Cry 5 is still slightly behind the GeForce GTX 1070 Max-Q in the Razer Blade 15.

More distance to the desktop GPUs

If desktop GPUs are included in the UHD performance rating (less CPU limit), the GeForce RTX 2070 Max-Q in the "cold" state in the Gigabyte Aero 15-X9 is on average as strong as the GeForce GTX 1070 Founders Edition , However, the desktop graphics card has the advantage of a significantly stronger CPU, which spoils the comparison. Another statement, however, is less strong because it affects both perspectives: the GeForce GTX 1070 FE is on average 18 percent ahead of the GeForce GTX 1070 Max-Q, the GeForce RTX 2070 Turbo of the GeForce RTX 2070 Max-Q, however, by 37 Percent. The gap between desktop and notebook has widened significantly, at least based on the systems available so far.

The smaller Pascal GPU, on the other hand, fits perfectly into the scheme called for by Nvidia, which says: Mobile GPUs are a maximum of 10 percent slower than the Founders Edition and Max-Q variants are a maximum of 10 percent slower than the conventional mobile versions. So far, there is no such statement at Turing.

On the next page: Performance with "RTX on"


Notice: ob_end_flush(): failed to send buffer of zlib output compression (1) in /home/gamefeve/bitcoinminershashrate.com/wp-includes/functions.php on line 5373

Notice: ob_end_flush(): failed to send buffer of zlib output compression (1) in /home/gamefeve/bitcoinminershashrate.com/wp-includes/functions.php on line 5373