RTX 5090D tested with nine-years-old Xeon CPU that cost $7 — it does surprisingly well in some games, if you enable MFG
February 7, 2025

RTX 5090D tested with nine-years-old Xeon CPU that cost $7 — it does surprisingly well in some games, if you enable MFG

The Chinese reviewer of technical technologies supported the completely new “for China” RTX 5090d NVIDIA, from the flagship AMD Ryzen 7 9800x3D, one of the best processors for games, to the Core Ultra 9 285K Intel. The four-core Xeon E5-2680 V4 aged nine years based on the Broadwell architecture. Despite age, the chip for $ 7 was at least moderately competitive with modern processors when using DLSS frames, especially with DLSS4 MFG.

The reviewer checked Counter-attend 2IN Marvel competitorsAnd Cyberpank 2077 On the assortment of processors paired with RTX 5090D. The list of processors consists of Core Ultra 9 285k, Core Ultra 7 265k, Core Ultra 5 245k, Core i7-14700KF, Core i5-14600KF, Core i5-12400F, Core i3-12100F and XEON E5-2680 V4. On the side of AMD, Ryzen 99900x, Ryzen 7 9800x3D, Ryzen 5 9600x, Ryzen 9 7900x, Ryzen 7 7800x3D, Ryzen 5 7500f and Ryzen 5 5600 were tested. This is a wide range of processors extending for almost ten years.

Counter-attend 2 showed the lowest results for Xeon E5-2680 V4. With maximum 4K settings, the average frequency of the CPU personnel was 33% slower than the next Core i3-12100F, not to mention higher performance chips such as Ryzen 7 9800x3D, which were almost twice as fast. A Broadwell chip provided the average frame rate of 168 frames per second, and the Core i3-12100F-248 frames per second. The fastest processors exceed up to 322 frames per second – all without any kind of frame generation. 1% with a low frame was even worse: the Xeon E5-2680 V4 grid is only 72 frames per second, and the fastest planting of the CPU at a speed of 163 frames per second is 2.26 times faster.

The terrible frequency of the Xeon Chip personnel CS2 Shows a problem with a combination of such a powerful graphic processor with an “archaic” processor. But this testing RTX 5090D also serves as the basis for other tests.

Marvel competitors With 4K settings of the highest quality, it almost completely changes CS2 The results are all processors from the I3-12100F nucleus to the Core Ultra 9 285K, exchanged almost identical frame rate of 111-115 frames per second, which detects a narrow part that is not related to the processor. Xeon E5-2680 V4 was not far behind, with an average frequency of frames 101. This is done by the fastest Ryzen 7 9800x3D only 14% faster than the nine-year Xeon chip. But again, the minimum FPS is important, and Xeon controlled only 65 frames per second compared to 80–99 frames per second on other processors. This is 34% slower at the minimum frame.

It is interesting that the DLSS 3 personnel generation is more than doubled by the performance of all processors, while most chips are at the level of 238–242 frames per second. (We must assume that DLSS also turned on, because otherwise performance should not be more than 80 ~ 90 percent faster, at best from Framegen.) With DLSS3 Framegen, Xeon E5-2680 V4 slightly closed the gap, on average on average 218 frames per second. However, the minimums were still 38% slower.

And DLSS4 with MFG4X in the competitors of Marvel pushed everything into a range of more than 400 FPS on average. This time, Xon was only 6% behind the fastest chips. And, of course, 1% of the minimum was still a little worse, at a speed of 164 frames per second compared to 298–308 a frame per second on the fastest processors. Thus, this is still a significant deficit of 40–47 percent, and, as we indicated in our testing DLSS4 and MFG, when you divide by four to get the basic personnel and input speed, it worked only at 41 frames per second before Framegen.

Cyberpunk 2077, With the maximum RT-overheating (possibly, psycho), the settings were only 34 frames for the best processors, including Ryzen 7 9800x3D. Slow chips, such as Ryzen 5 5600 and Core i3-12100F, landed at a speed of 32 and 31 frames per second, respectively. Xeon E5-2680 V4 was again the lowest, but still with a middle frame of 29 frames per second. In other words, Ryzen 7 9800x3D surpassed Xeon E5-2680 V4 by 17%. The minimum FPS was 19 on Xeon versus 24 on other processors, 26%.

DLSS3 Framegen and increased scaling give a huge pulse for personnel, and planting Xeon at a speed of 95 frames per second compared to 125–132 frames for other chips. In this case, it was 28% slower. But then DLSS 4 with Multi Frame Generation (MFG) increases the frame even more.

Ryzen 7 9800x3D controlled 255 frames per second, about 250 frames per second on all other AMD chips, with the exception of Ryzen 5 5600. DLSS 4 with MFG4X on Xeon E5-2680 V4 allowed the ancient CPU almost coincide with other Intel chips on average FPS V4 . He scored 213 frames per second compared to 219–226 frames per second on other Intel processors. This is only a difference of 6%, most … If you do not look at the minimum FPS. In this case, Xeon controlled only 78 FPS (mainly binding 12100f), while most other processors were in the range of 120-150 frames per second.

In short, no, the Xeon nine -year -old processor will not keep up with the last chips, with or without Framegen, MFG and scaling. But this does better than you could expect at least three games that were tested on RTX 5090D. Personnel generation may be especially help in games that are fully limited by CPU, as this means that the graphic processor has many spare ones to launch the artificial intelligence procedure to generate personnel. This is especially useful for Xeon E5-2680 V4, as it will be a very limited CPU.

This does not mean that MFG and Framegen are a solution for narrow places of the processor. If you take some of the minimum FPS results and divide by four (for example, in Cyberpunk 2077), the base frame was less than 20 frames per second. Technically, this can be reproducible after four times up to 78 frames per second, but this is not the most smooth experience at all.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *