Again, strange results in GTA V (2015) - only 138.5fps in QHD and for some reason 175.7fps in 4k, which requires much more resources. All other games are exactly the opposite. There is a clear error from test to test from review to review.
What do we see in comparison with Razer on the same processor?
PL1=150W in Titan or 37% higher than Razer. The difference in speed is only 6% in CBR15, and judging by the gradual drop in numbers, you had to continue the test for another 30 minutes. Especially for Titan with a separate extended schedule. It is obvious that the laptop cannot keep these 4900 parrots further and the computing speed will gradually drop. Clearly extremely inefficient use of energy in the MSI Titan.
Go ahead...
Razer with the same processor has a much faster 5600 memory than the infamous 4800. What immediately affects the AIDA64 tests for working in Photoshop is that Razer is almost 25% faster in the Photoworx test. A very significant difference for a professional. And all because of the slow 4800 memory, which have throughput only a shameful 60-63Gbyte/s, instead of 78-83Gbyte/s in Razer with 5600.
Let me remind you again that the memory controller in Apple's M2 Max is almost 5 times faster than that of the i9 13950HX - it pumps up to 400GByte/s. Camp x86 is completely defeated. Years ahead...
Let me remind you again that the memory controller in Apple's M2 Max is almost 5 times faster than that of the i9 13950HX - it pumps up to 400GByte/s. Camp x86 is completely defeated. Years ahead... For example, in Zen4 7945HX there is simply nothing to serve 28 pci-e 5.0 lines, which simultaneously require at least 112 GByte/s, and the memory speed is much lower, but should be (for everything else to work easily) up to 10 times higher, i.e. almost 1TByte/s for top mobile Zen4. When will we see a 512-1024 bit memory controller in the x86 camp? With pumping from 0.5TByte/s at least for RAM speed? We don't need a 100MB L3 cache (like SLC cache in SSD). We need HBM memory in x86. Here and now...
System latency:
Despite the wild CPU and super speed, it has obvious problems with smooth video and sound, according to Latency Monitor tests. MSI needs to do something urgently with the drivers. However, Razer is also not doing well in this matter.
The main thing in 3D is that the games finally got stable 60fps + with a mobile 4090. But of course, there is no question of any 120Hz +. But thanks to a versatile 16:10 4k panel that easily switches to fhd mode for crystal-clear imagery (as opposed to Razer's hazy 2.5k fhd mode), they'll easily get over 100fps in any game at Ultra settings, if desired.
Key notes about 4k miniLED panel:
1. A key note about 144Hz mode for this panel is that to support HDR operation at this frame rate, the eDP link must be pumping around 40Gbps. But as far as I know, neither RaptorLake nor Zen4 (even Phoenix with "4nm") support UHBR20 mode. Zen4 officially only supports UHBR10 in the Phoenix version, and Raptor Lake generally only infamous eDP1.4b with a maximum of 32Gb / s (including overhead and 25.92Gb / s without). As everyone knows, officially NVidia and AMD also do not support DP2.0 (2.1) in their video chips for laptops in 2023, and certainly not in UHBR20 mode. It turns out that MSI forgot to inform owners of Titan laptops that the panel is connected to a laptop via an interface that is not able to transmit a 3840x2400@144Hz 30bit (HDR) signal in lossless mode. And transmits it only in lossy DSC mode, i.e. the picture is not a perfect monitor lossless mode 4:4:4. Thus, even the 120Hz mode in HDR is not achievable here, if there is eDP1.4b in the i9 only and only with DP2.0+ with UHBR10 120Hz with HDR are achievable...
2. The author took measurements contrast of miniLED in native mode, knowing full well about the key problem of miniLED - halos around text at the borders of light and dark zones. This is an irreparable shortcoming. Is it possible to turn off multi-zone backlighting on this panel via MSI software and work with one common backlight, as in some miniLED monitors? If there is, I would like to see tests of native contrast with multi-zone highlighting turned off and without halos around the text. Only microLED will give us an analogue of AMIOLED, but on the condition that it will not flicker and will have a resource of 15k hours, as opposed to 7-8k hours on a typical AMOLED, all other things being equal.
So far, the usual IPS 4k@120Hz (in 24-bit mode) with a native contrast ratio of 1500:1+ remains the best choice for the eyes.
3. The response time cannot be zero a priori.
The fact that the laptop is very noisy (traditionally for MSI) is obvious and there is no point in focusing on this. Although I would like to get a more detailed study of the laptop's operation in a normal load when setting up the operation of coolers through MSI's proprietary cooler management utility - this is their significant advantage over laptops from other companies, only L7Pro 2023 supposedly will receive the same utility for the first time in Lenovo's history - please and there to investigate this question carefully when you have it on tests.
What can I say in the end, given the weight of the laptop along 5kg+ with PSU and its wild noise in games? What is this all for? This is no longer a laptop, but a portable system unit with a built-in monitor. In reality, only shift workers and sailors at sea need such a machine, because they just have nothing else to do there. But why somewhere need to buy this monster for home, instead of a much quieter desktop computer with a much faster desktop 4090?
Deutsch
English
Español
Français
Italiano
Nederlands
Polski
Português
Русский
Türkçe
Svenska
Chinese
Magyar