AMD Ryzen 9 7900X and Ryzen 5 7600X, reviews

The Ryzen 9 7900X and Ryzen 7 7600X are two of the most anticipated processors in the Ryzen 7000 series , whose presentation took place on August 30. It was undoubtedly one of the most important events in the technology sector, and it is perfectly understandable, not in vain we speak of a generation that picks up the witness of the Ryzen 5000 , and that has to face two important challenges : overcome the bar that they have left these, and frankly it has been very high, and show that they have what it takes to lead the field of high-performance general consumption x86 CPUs.

I am not going to lie to you, after writing so much about the Ryzen 7000 I really wanted to be able to try them, and thanks to AMD Spain I have been lucky enough to receive not one but two units, the Ryzen 9 7900X, which is a high- end model , and the Ryzen 5 7600X , which positions itself directly in the mid- range and succeeds the legendary Ryzen 5 5600X, one of the best gaming processors AMD has launched so far.

AMD has completely changed the design of the box in the Ryzen 7000

At first glance we might think that there are no major changes, since the Ryzen 9 7900X has 12 cores and 24 threads, the same as the Ryzen 9 5900X, and the Ryzen 5 7600X has 6 cores and 12 threads, the same amount as the Ryzen 5 5600X, but nothing could be further from the truth. Both processors bring profound changes at the microarchitecture level , are manufactured in a more advanced node and are integrated into a latest generation platform.

The changes introduced by the Zen 4 architecture have made possible an increase in the IPC which, according to AMD itself, is 13% compared to the Ryzen 5000. This means that a Zen 4 core running at the same working frequency as a core Zen 3 performs 13% more. However, if we add to this the increase in working frequencies, we see that single-thread performance has increased by 29% compared to the Ryzen 5000. Impressive, without a doubt.

AMD has kept the MCM design in the Ryzen 7000, which means that the CCD unit is once again the central pillar and the I/O chip is still present. The first one is made in 5nm node , and the second one uses 6nm node . TSMC has been the “cook” for both, as our regular readers will remember, and all those chiplets are interconnected through the Infinity Fabric system.

Ryzen 9 7900X frente a Ryzen 9 5900X

With this comparison we are going to discover in a direct and simple way what are the key differences that exist between both processors. We are not going to delve into details at the level of architecture or performance, since we will talk about it later, we will only see its specifications to have a first contact with its characteristics and differences.

Ryzen 9 5900X

• Zen 3 architecture, manufactured on TSMC’s 7nm node.
• Two CCD units with 8 cores each and SMT technology, which allows handling one process and one thread per core.
• 12 cores and 24 threads at 3.7GHz-4.8GHz, normal and turbo mode.
• 64 MB of L3 cache (32 MB per CCD unit) and 6 MB of L2 cache (512 KB per core).
• 105 watt TDP.
• Compatible with DDR4 and PCIe Gen4 memory.
• Support overclock.
• It uses the AM4 socket, and is compatible with motherboards equipped with the 300, 400, and 500 series chipsets.
• It does not have an integrated GPU.

Ryzen 9 7900X

• Zen 4 architecture, manufactured on TSMC’s 5nm node.
• Two CCD units with 8 cores each and SMT technology, which allows handling one process and one thread per core.
• 12 cores and 24 threads at 4.7 GHz-5.6 GHz, normal and turbo mode.
• 64 MB of L3 cache (32 MB per CCD unit) and 12 MB of L2 cache (1 MB per core).
• 170 watt TDP.
• Compatible with DDR5 and PCIe Gen5 memory.
• Support overclock.
• It uses the AM5 socket, and is compatible with motherboards equipped with the 600 series chipsets.
• AVX512 instructions.
• It has an integrated Radeon RDNA2 GPU with two compute units (128 shaders).

The differences between both processors are notable, and are clear at a glance. The Ryzen 9 7900X uses a more advanced node, supports next-generation technologies and standards , comes with an integrated GPU , which gives it more value at a functional level, and increases the CPU’s working speed by 1 GHz in base mode. and at 800 MHz in turbo mode, two impressive values ​​explain that great improvement in single-thread performance that we referred to, and that show that AMD’s chiplet design has matured like fine wine.

The use of DDR5 memory will also represent a significant increase in performance, in fact we have already begun to see it in some games, such as Spider-Man Remastered, for example, and compatibility with the PCIe Gen5 standard will allow us to enjoy a wide higher bandwidth both with graphics cards and storage units compatible with said standard.

We can also see one of the most important changes at the microarchitecture level that Zen 4 has introduced, the increase in the L2 cache , which becomes 1 MB per core, and we must not forget that the jump to the 5nm node has allowed AMD to reduce the size of the transistors. This has made it possible for CCD units to have higher transistor density despite being smaller than Zen 3.

Ryzen 5 7600X frente a Ryzen 5 5600X

We now turn to compare these two mid-range stars. Both have a very similar configuration, but as happened with their older brothers, it is enough to put their specifications face to face to realize that the differences they present are very important. As in the previous case, we are not going to delve into details at the architecture level, we will leave that for later.

Ryzen 5 5600X

• Zen 3 architecture, manufactured on TSMC’s 7nm node.
• A CCD unit with 6 active cores and SMT technology, which allows handling one process and one thread per core.
• 6 cores and 12 threads at 3.7GHz-4.6GHz, normal and turbo mode.
• 32 MB of L3 cache (32 MB per CCD unit) and 3 MB of L2 cache (512 KB per core).
• 65 watt TDP.
• Compatible with DDR4 and PCIe Gen4 memory.
• Support overclock.
• It uses the AM4 socket, and is compatible with motherboards equipped with the 300, 400, and 500 series chipsets.
• It does not have an integrated GPU.

Ryzen 5 7600X

• Zen 4 architecture, manufactured on TSMC’s 5nm node.
• A CCD unit with 6 active cores and SMT technology, which allows handling one process and one thread per core.
• 6 cores and 12 threads at 4.7GHz-5.3GHz, normal and turbo mode.
• 32 MB of L3 cache (32 MB per CCD unit) and 6 MB of L2 cache (1 MB per core).
• 105 watt TDP.
• Compatible with DDR5 and PCIe Gen5 memory.
• Support overclock.
• It uses the AM5 socket, and is compatible with motherboards equipped with the 600 series chipsets.
• AVX512 instructions.
• It has an integrated Radeon RDNA2 GPU with two compute units (128 shaders).

As in the previous comparison, the differences are very clear. The Ryzen 5 7600X has the same number of cores and threads as the Ryzen 5 5600X, but runs at much higher frequencies ( up to 1 GHz more ), has double the L2 cache , is compatible with the latest generation standards and has a Integrated GPU, giving it greater functional value.

I know that some of our readers will be wondering why we place so much value on the integrated GPU, and it’s very simple, because it can get us out of trouble if the dedicated GPU breaks down, and because in this case AMD has not had to make sacrifices to be able to integrate it, although we will talk about it later.

Zen 4 in depth: analysis of the architecture of the Ryzen 9 7900X and Ryzen 5 7600X

The Zen 4 architecture maintains the legacy of the Zen 3 architecture , a fact that can be seen in three main keys: the use of the CCD unit (also known as a chiplet) as a central pillar, the maintenance of the I/O chip and the system of Infinity Fabric interconnection to communicate both elements.

Although the base of Zen 4 is the same as that found in Zen 3, AMD has introduced important innovations that we will analyze in depth below. The first important change is in the increase in the CPI which, as we have already indicated at the beginning of the analysis, has grown by 13%. We have already explained to you what this means, that a Zen 4 core is 13% faster than a Zen 3 running at the same frequency, but how has AMD achieved such a big IPC improvement? Let’s find out by reviewing the most important changes:

• Improvements in the predictor of jumps, in the front end and in the loading and storage of data.
• Increased L2 cache, going up from 512KB to 1MB per core.
• Deeper buffers through the core.
• Largest Int/FP log file.
• Larger Op Cache.
• Execution engine improvements.
• Enhanced Infinity Fabric interconnect system.

We must add to all this the support of AVX512 instructions through floating point units, an important advance that according to AMD does not affect the working frequencies of Ryzen 7000 processors, that is, in theory, working with AVX512 does not force reduce CPU speed. These instructions make a big difference with workloads focused on inference, deep learning, and AI.

The CCD unit has been maintained, but it has changed, and not only because of those novelties introduced by the Zen 4 architecture, it has also done so because of the jump from the 7nm to the 5nm node. Thanks to this new process, AMD has been able to reduce the surface area of ​​a Zen 4-based CCD unit to 70 square mm , a remarkable reduction when compared to the 83 square mm of a Zen 3-based CCD unit.

A Zen 4 chiplet is smaller than one based on Zen 3, but at the same time it has a larger number of transistors. This is possible because when moving to the 5nm node the size of the transistors is considerably reduced, and this allows a higher density to be integrated in a smaller space. For comparison purposes, the Zen 4 CCD used in the Ryzen 7000 has 6.57 billion transistors , while its Zen 3 equivalent has 4.15 billion transistors.

Keep in mind that we are talking about density of transistors per CCD unit , which means that those processors that use two CCD units with all their cores active, such as the Ryzen 9 7950X, will have a density of 13.14 billion transistors (the sum of both units). The 5nm node, and the changes that AMD has introduced at the architectural level, have also made it possible to greatly increase working frequencies compared to the previous generation, and increase energy efficiency by 40% on average (performance per watt consumed).

The improvement of the working frequencies deserves a separate mention, and that is that what AMD has achieved with the Ryzen 7000 has been something very big. I still remember when we questioned whether it was going to be able to overcome the 5 GHz barrier with an MCM architecture, and today, a couple of years later, we can say that it has not only overcome it, but has even approached 6 GHz. GHz, since the Ryzen 9 7590X has a turbo mode that, under the right conditions, can reach 5.85 GHz. If we compare with the Ryzen 5000, the frequency differences can reach 1 GHz, impressive considering that it is it is a high that has occurred from one generation to another, and not between several generations.

The turbo mode allows the working frequencies to scale dynamically depending on the workload, and in general it acts in the same way as in the previous generation, although in the Ryzen 7000 it is capable of reaching much higher levels. . For example, a Ryzen 9 7950X works at 4.5 GHz in normal mode, but can reach 5.7 GHz in turbo mode , and peaks at 5.85 GHz if Precision Boost Overdrive is activated. Well, these frequencies are the maximum level possible with a certain number of active cores and threads, that is, the greater the number of these active, the lower the maximum frequency of turbo mode.

Continuing with the previous example, we see that the Ryzen 9 7950X drops to 5.6 GHz as soon as we have three active threads, and from there it maintains a gradual drop that takes it to 5.21 GHz maximum frequency with 32 active threads . What can I say, it’s not bad at all, as it continues to far exceed the maximum 4.4 GHz that the Ryzen 9 5950X was capable of maintaining with all its cores and threads active.

Zen 4 has not increased the maximum cores and threads over Zen 3 as each CCD unit maintains the 8 cores and 16 threads as well as the 32MB L3 cache configuration. There is also no change in access to said cache , which means that each of those 8 cores can access the 32 MB of L3 cache. However, thanks to the IPC improvements that we have indicated, the increase in the L2 cache and the great rise in working frequencies, the performance difference between Zen 3 and Zen 4 is up to 29% per core in favor of the second. For comparison purposes, AMD ensures that Zen 4 has improved the IPC by 242% against Piledriver , the veteran AMD FX.

We have already said that AMD has managed to increase the density of transistors despite having reduced the size of the CCD unit, and that the working speeds have increased impressively. These are two important achievements that, however, have had a notable impact on consumption and working temperatures . By having more transistors concentrated on a smaller surface, the heat accumulates more quickly, and to maintain such high working frequencies, AMD has had to increase the TDP levels. This is obvious in the comparison that we have left you, since the Ryzen 9 7900X has a TDP of 170 watts compared to the 105 watts of the Ryzen 9 5900X, and the Ryzen 5 7600X has a TDP of 105 watts compared to 65 watts of the Ryzen 5 5600X.

It’s a problem? Well, the truth is that no, it is totally understandable that AMD has had to increase consumption to maintain such high working frequencies, and as we have said, we are talking about an increase of between 800 MHz and 1 GHz, depending on the values ​​at the ones we look at In terms of transistor density and heat, the Sunnyvale firm has used gold plating on the CCD units to improve heat transfer, and has repeated the use of solder on the IHS, a show of good work on the part of from AMD.

We now jump to talk about the I/O chip, a solution that this time comes with an important novelty, and that is that it not only incorporates all the input and output elements, including the PCIe Gen5 and DDR5 controllers , but this time also it comes with an integrated GPU. This has a total of two compute units, and each of them has 64 shaders, 4 texturing units, and one ray tracing acceleration unit, leaving us with a total of 128 shaders, 8 texturing units, and two units to speed up ray tracing. This GPU uses the RDNA2 architecture , supports AV1 decoding , HDMI 2.1 and DisplayPort connectivity and can also work with the H.264 and H.265 standards.

The I/O chip is manufactured on TSMC’s 6nm node , which means that in a Ryzen 7000 two manufacturing nodes coexist, this one and the 5nm one used in the CCD units. The communication between all these chips is done through an Infinity Fabric system that has its own clock frequency. In this new generation of processors, the Infinity Fabric Clock (FCLK) works by default at 1,733 MHz, and AMD itself has confirmed that the ideal is to let it work in automatic mode to achieve the best performance, that is, so that everything works in 1:1 mode .

This is very important, because if the balance were to be broken between the Infinity Fabric Clock, which determines the speed at which the CPU cores can communicate with the other chiplets, the Memory Controller (UCLK), which refers to the maximum speed at which the memory controller can work, and the Memory Clock (MCLK), the speed of the RAM memory that we have installed, we could enter 1:2 mode , which would cause the speeds to be reduced by half, and this would cause a significant performance loss.

To maintain that balance, AMD has confirmed that 6 GHz DDR5 memory is ideal,  as the system will automatically adjust to 3000 MHz (MCLK):3000 MHz (UCLK):2000 MHz (FCLK), i.e. in mode 1 :1. If the memory frequency (MCLK) rises above 3,000 MHz (6,000 MHz because it is double data rate) the 1:2 mode will be activated. Please note that the MCLK and FCLK ratio will always be 3:2. Indeed, in the Ryzen 7000 it is not necessary to maintain the 1:1:1 mode to achieve optimal performance.

AMD AM5 vs. AM4: this is the new platform used by the Ryzen 9 7900X and Ryzen 5 7600X

With the arrival of the Ryzen 7000 processors, AMD launches a new platform known as AM5 , a name that is no coincidence as it clearly indicates that we are facing the successor of AM4, and that in general it deserves to be considered as a new generation , since it is prepared to work with DDR5 memories at high frequencies and is also compatible with the PCIe Gen5 standard.

Before talking about the most important news that this platform brings, and the chipset that we have used in our tests, the X670E, it is important to touch on another topic, the socket, and that is that with the Ryzen 7000 there has been another important change. in this sense. As many of our readers will know, the AM4 platform uses a PGA-ZIF type socket with a total of 1331 contacts . This means that it is prepared to use processors with arrays of pin grids , which are inserted into the socket and are held in place by the pressure exerted by the socket’s closing mechanism.

The AM5 socket, also known as LGA-1718 due to the number of contacts , is very different, since it changes the PGA-ZIF interface to an LGA-ZIF type. This means that it no longer supports processors with pin arrays, and is now compatible with models that use a grid contact array , the same contact mechanism that Intel has been using in its processors for many years. The ZIF system is maintained, which means that it is not necessary to apply force to place the processor, and that it is held by the pressure applied by the socket closure.

Here we can see the LGA contacts of the new Ryzen 7000

Another important change that applies to the socket is that it can now offer up to 230 watts of power to the processor (PPT), something that obviously improves performance but has a considerable impact on consumption and working temperatures, and that will be unnecessary. if the chip has already peaked at the temperature level. However, it must be taken into account that this maximum does not apply equally to all processors, and that it is divided as follows:

• Ryzen 7000s with a TDP of 65 watts can reach a PPT of 88 watts.
• Ryzen 7000s with a TDP of 105 watts can reach a PPT of 142 watts.
• Ryzen 7000s with a TDP of 170 watts can reach a PPT of 230 watts.

We continue to delve into other important changes within the AM5 platform versus the AM4 platform, and we find an increase in PCIe lanes, since we can count on up to 20 PCIe Gen5 lanes and up to 44 PCIe lanes in total , support of a greater variety of ports and other improvements to the level of sound, bus and power control system of the platform. In the attached image you can see a complete breakdown with all the most important keys of the X670E chipset which, we remember, is dual , that is why an “upstream” and a “downstream” unit (upstream and downstream) are differentiated.

It is very interesting that AMD has decided to use a dual chiplet on the AM5 platform, although we must remember that this is limited to the high end, that is, the X670E and X670 series, since the motherboards equipped with the B650E and B650 use a single chipset . For comparative purposes I wanted to share with you an image in which we can see all the keys of these chipsets.

There is no doubt that the evolution represented by AM5 compared to AM4 is very great, so much so that it represents the entry of AMD into the new generation , championed mainly by support for DDR5 memory at 5,200 MHz and compatibility with the PCIe Gen5 standard. Keep in mind that even if the natively supported memory is at a frequency of 5,200 MHz, we can use configurations at a higher frequency without problems, thanks to AMD EXPO technology , which, broadly speaking, is a version of Intel XMP technology.

Click to enlarge

With AMD EXPO we can apply overclocked memory profiles and tuned latencies in a safe and simple way. We just have to enter the BIOS and choose the profile we want to activate, without further ado. In this review we have actually used 6,000 MHz DDR5 memory and we have not had any problems . Booting through the BIOS was as simple as a lollipop mechanism, and ran smoothly at that frequency.

Test and configuration team

To analyze the Ryzen 9 7900X and Ryzen 5 7600X we have used a configuration at the level of both processors, consisting only of high-end components and accompanied by Windows 11 updated to the latest version available . Here is a list of all the components we have used.

• Ryzen 9 7900X 12 core 24 thread 4.7GHz-5.6GHz processor, normal and turbo mode.
• Processor Ryzen 5 7600X 6 cores and 12 threads at 4.7GHz-5.3GHz, normal and turbo mode.
• 32 GB of DDR5 G.Skill Trident Z5 Neo RAM at 6,000 MHz with AMD EXPO technology and CL30 latencies configured in dual channel (two modules of 16 GB each).
• AORUS Master X670E baseboard.
• Corsair iCUE H150i Elite LCD All-in-One Liquid Cooling System with Three 120mm Fans.
• 480GB Crucial SSD.
• WD Black SN850 2TB SSD with PCIe Gen4 x4 interface, capable of speeds of 7,000 MB/s and 5,300 MB/s in sequential read and write.
• GeForce RTX 3090 Ti Founders Edition Graphics Card.
• Corsair RM1000x 80 Plus Gold certified power supply with official NVIDIA 12-pin adapter cable.
• Windows 11 updated to the latest version available.
• All drivers and drivers updated to their latest versions.

GeForce RTX 3090 Ti Specifications

• GA102 GPU manufactured in 8nm process.
• 10,752 shaders at 1,560 MHz-1,860 MHz, normal and turbo mode.
• 336 textured units.
• 112 units of raster.
• 336 tensor cores.
• 84 cores RT.
• Bus de 384 bits.
• 24 GB of GDDR6X memory at 21 Gbps.
• It uses the PCIe Gen4 standard in x16 mode.
• Bandwidth of 1,008 GB/s.
• Power in FP16: 40 TFLOPs.
• Power in FP32: 40 TFLOPs.
• Power in FP64: 625 GFLOPs.
• TGP: 450 watts.

In the tests we have also used AMD’s Ryzen Master tool . With it we have been able to measure important aspects such as working frequencies, consumption, voltage and amperage. This tool also allows us to overclock manually or use PBO technology, also known as Precision Boost Overdrive , which can automatically increase working frequencies depending on certain aspects, such as the number of active cores and threads, consumption and working temperatures.

Preliminary considerations on the assembly and stability test

I already told you at the time when we unpacked some of the most important “secrets” of the new Ryzen 7000 processors that the “octopus”-shaped IHS was no coincidence . That heat spreader has an explanation, and that is that it has been essential for AMD to maintain full compatibility with socket AM4 cooling systems on socket AM5.

We didn’t need any kind of adapter, and mounting the cooling system was no mystery. Regarding the issue of applying thermal paste, I can confirm that the rumors were true, you have to apply less quantity and carefully, because if we exceed it, it will overflow.

In general, with a small ball of the size that you see in the image is more than enough, in fact I would recommend a little less since when I changed from the Ryzen 9 7900X to the Ryzen 5 7600X I saw that it had overflowed a little on two sides, nothing serious but it’s annoying to clean. As always, I recommend that you opt for a metal-free paste, that is, non-conductive , since in case you stain the base plate or the baseboard, you will have nothing to fear. If you are curious, I can confirm that in this analysis I have used Arctic MX-4 thermal paste, which is the same one that I use on my personal PC and that has always given me very good results.

I have used a 360mm AIO liquid cooling kit because the Ryzen 9 7900X and Ryzen 5 7600X come without a stock fan, and because this is the one that AMD openly recommends for the Ryzen 9 7900X . With the Ryzen 5 7600X we would not have any problem if we used a more modest cooling system, as we will see later when we talk about working temperatures.

During the assembly process I had no problems , and the process of updating the motherboard BIOS, which came with a very early version, was completed in just a few minutes. It is true that in the first full boot the test equipment took longer than usual for a PC of its characteristics, but AMD had already warned me that it was normal, and indeed it was. After the first boot I rebooted into the BIOS and loaded the AMD EXPO profile to bump the RAM up to 6GHz.

On my second arrival at the Windows 11 desktop I wanted to run a stability test to see if everything was in order. For this I used the AIDA64 stress test, and everything went smoothly, I did not register any type of problem in the more than 30 minutes that I left said test activated. Since everything was in order, it was time to start passing performance tests, and the first processor I used was the most powerful, the Ryzen 9 7900X.

Ryzen 9 7900X: Synthetic Test Performance

We start with a classic, CPU-Z. In this we can see that the Ryzen 9 7900X reaches almost 772 points in single thread , a figure that places it only slightly below the Intel Core i9-12900K , which scored 800 points in our analysis. In multithreading, the AMD processor reaches 11,558 points, a figure that places it almost at the level of the Ryzen 9 5950X , impressive since the latter has 16 cores and 32 threads, while the Ryzen 9 7900X adds 12 cores and 24 threads. Compared to the Intel chip, it achieves a narrow victory, since it achieved 11,255 points in multithreading.

We now turn to see the data from Cinebench R23. In single thread we have an impressive score, 1,997 points . The Ryzen 9 7900X touches the 2,000 point barrier and clearly outperforms the Intel Core i9-12900K, which scored 1,888 single-threaded. In multithreading, the AMD chip achieves 27,190 points , a figure that also allows it to surpass the Intel chip, since it obtained 26,883 points. The thing is adjusted, but the numbers do not lie.

We go to see the results of PassMark and we have a result that lives up to expectations, 52,647 points , a figure that also exceeds the Intel Core i9-12900K, since it obtained 43,940 points , and that places the Ryzen 9 7900X as one of the most powerful processors in the world. It is a simple test, but quite complete and very useful for reference.

In the attached graph you can see a relative performance comparison between the Ryzen 9 7900X and the Core i9-12900K in Cinebench R23 , both running at stock frequencies and with the same cooling solution, a Corsair iCUE H150i Elite LCD AIO kit with three 120mm fans. AMD’s processor wins by more than 5% in single threading and by just over 1% in multithreading.

The performance achieved by the Ryzen 9 7900X in V-Ray is also fantastic, nothing more and nothing less than 22,234 “vsamples”, a figure that exceeds the 19,982 “vsamples” that the Ryzen 9 5950X obtains , and also almost 18,400 “vsamples” that gets the Intel Core i9-12900K. It is clear that AMD has done a good job, and that it has managed to significantly increase the raw performance of its processors with the Zen 4 architecture. It was not bluffing when it spoke of an increase in IPC of 13% compared to the Ryzen 5000, and an increase of up to 29% in single-wire performance.

In the AIDA64 FPU AES performance test we also have an excellent result, since as we can see in the attached image the Ryzen 9 7900X is in second position , just a little below a third generation Ryzen Threadripper with 32 cores at 3,800 MHz. We also see that it easily exceeds the Core i9-12900K and the Ryzen 9 3950X, which adds 16 cores and 32 threads.

We continue with AIDA64, and this time we go to the memory and cache test. The results we have obtained are also very positive , since AMD has fine-tuned the latencies at the L3 cache level and has improved its bandwidth. The values ​​that we have obtained in the L2 cache are slightly lower than those that we saw in the previous generation, but this may be a direct consequence of the increase in this, which, as we have already told you, has doubled when going from 512 KB to 1 MB per core.

Blender’s CPU test also shows that the Ryzen 9 7900X is a high-performance processor, and that it is ready to handle heavy workloads in professional applications without any issues. It’s not just a high-performance gaming processor, it’s so much more.

In 3DMark CPU we can see some excellent performance data , quite contained working temperatures that vary depending on the frequency and the number of active threads, and a speed scaling that fits perfectly with the values ​​promoted by AMD.

• With an active thread, the frequency reaches 5,698 MHz and the temperature stabilizes in the range of 72 degrees.
• With two active threadswe have a fairly stable frequency of 5,670 MHz and a temperature of 74.75 degrees.
• With four active threadswe have temperatures of almost 76 degrees and a frequency of 5,515 MHz.
• With eight active threads, the frequency drops to 5,420 MHz and temperatures reach 80.63 degrees.
• With all threads active, the temperature reaches 83.75 degrees and the speed remains stable at 5,344 MHz.

You can enlarge the gallery below by clicking on it to see the results in more detail.

Ryzen 9 7900X: Gaming Performance

Synthetic performance data confirms that the Ryzen 9 7900X represents a significant leap from the previous generation , quite an achievement as the Ryzen 5000 became the most powerful processors of its generation, both multi-core and single-core thanks to its high IPC, which even allowed them to compensate for the fact that they worked at a lower frequency than the Core Gen11, which were their direct rivals.

We now turn to see the performance in games. I have included results in 720p and 1080p because these low resolutions are the ones that put the most pressure on the processor , and in which it is easier to see differences between different generations. This has an explanation, and it is that in such low resolutions a graphics card like the GeForce RTX 3090 Ti goes so far that it needs the CPU to provide it with the greatest possible amount of data and instructions, while in 4K the opposite happens, since that it is the GPU that assumes the greatest load and the CPU is much more loose.

In Red Dead Redemption 2 we already see that there is a substantial difference compared to the Ryzen 5000. In 1080p a Ryzen 9 5950X achieves an average of 101 FPS, while the Ryzen 9 7900X reaches 117 FPS, which leaves us a difference of 15 .84% in favor of the second . The differences are reduced as we increase the resolution, something that is totally understandable from what we have explained previously. At 1440p the Ryzen 9 7900X achieves 98 FPS, while the Ryzen 9 5950X reaches 92 FPS, which is a performance improvement of 4.26% . In 4K the difference is practically zero, something that is also totally normal.

Shadow of the Tomb Raider is one of the best games to test the performance of a processor because of its high dependency on the CPU, and because it scales better than others on processors with more than four cores and eight threads. With the Ryzen 9 7900X I have obtained excellent results in general, and the difference it makes compared to the Ryzen 9 5950X is quite large at low resolutions, although as we have seen before, that difference is reduced as the resolution drops.

At 720p the Ryzen 9 7900X achieves 262 FPS and the Ryzen 9 5950X achieves 206 FPS, which leaves us with an improvement of 27.18% . Going up to 1080p the difference between the two is reduced and the Ryzen 9 7900X achieves a performance improvement of 16.33% (228 FPS vs. 196 FPS for the Ryzen 9 5950X). At 1440p the gap between the two shrinks even further, with the Ryzen 9 7900X achieving 173 FPS and the Ryzen 9 5950X achieving 169 FPS. This leaves us with a victory of the first by 2.37%.

Cyberpunk 2077 is a game that also has a significant dependence on the CPU, especially the IPC and the working frequencies, something that can be seen in the results that we have obtained, as we will see just below.

With 1080p resolution, the Ryzen 9 7900X maintains an average of 152 FPS , a figure that far exceeds the 126 FPS that we obtain with a Ryzen 9 5950X, and which represents an improvement of 16.45%. By raising the resolution to 1440p, the Ryzen 9 7900X maintains a considerable advantage, as it achieves 105 FPS compared to 89 for the Ryzen 9 5950X, which represents a difference of 15.24% in favor of the first. In 2160p the difference is much smaller, but considerable, since the Ryzen 9 7900X maintains an average of 53 FPS and beats the 49 FPS of the Ryzen 9 5950X. In this case the difference is 7.55%.

Death Stranding is a game where it is very easy to generate a CPU bottleneck even at relatively high resolutions, and this makes it a very interesting option to analyze the performance of a processor. At 720p, a Ryzen 9 5950X achieves an average of 172 FPS, a figure far removed from the 228 FPS achieved by the Ryzen 9 7900X , and the GPU utilization rate at that resolution barely exceeds 50%. Raising the resolution to 1080p barely earns us a few FPS with the Ryzen 9 5950X, because it continues to produce a huge bottleneck, and this same trend remains at 1440p.

In Gears 5 the IPC of the Ryzen 9 7900X makes a substantial difference, but only at resolutions below 1440p . The truth is that this is the trend that I have seen in general during my tests, and it is something that we must take into account if we play in 1440p or 2160p, because at those levels the improvement is minimal or null.

The Ryzen 9 700X clocked in at 196 FPS at 720p, while the Ryzen 9 5950X clocked in at 132 FPS. The difference is an overwhelming 48.48%, a figure that confirms that with the second we have a big bottleneck at such a low resolution. By raising the resolution to 1080p, the performance with the Ryzen 9 5950X goes up to 136 FPS, a result that only confirms that bottleneck we talked about, and the Ryzen 9 7900X achieves 153 FPS, which translates into a victory for this by 12.5% . At 1440p we have practically a tie, as the Ryzen 9 5950X loses by only 1 FPS against the Ryzen 9 7900X.

We end with Metro Exodus Enhanced Edition, a game that I wanted to add to the comparison to give the analysis greater richness . I’ve used DLSS with different quality settings because it’s a realistic setting to counteract the impact of ray tracing, and because it lowers the base resolution and puts more weight on the renderer. I wanted to use the benchmark integrated in the game to better unify the results and simplify the comparison.

In 720p and 1080p, applying DLSS in quality mode reduces the pixel count so much that we have practically the same performance with the Ryzen 9 5950X, 129 FPS and 128 FPS, respectively. Going to 1440p with DLSS in Balanced mode that bottleneck is even more apparent as performance only drops to 125 FPS. For its part, the Ryzen 9 7900X scales much better and registers 159 FPS, 145 FPS and 134 FPS, giving it a clear victory. It renders 23.26% more in 720p, 13.28% more in 1080p and 7.2% more in 1440p.

To give you a clearer vision and facilitate the comparative analysis, I leave you the performance averages and the percentage difference between the Ryzen 9 7900X and the Ryzen 9 5950X. As we can see, the difference in 1080p is very marked, so much that it exceeds 17% , but this is reduced in 1440p to almost 8% in favor of the processor based on Zen 4 .

Ryzen 9 7900X: Temperatures and consumption

The Ryzen 9 7900X offers much higher performance compared to the previous generation from AMD. We have already seen that the jump in both single-wire and multi-wire is quite large, but this has had a considerable impact on both working temperatures and consumption.

Using a Corsair AIO liquid cooling kit the Ryzen 9 7900X scaled straight up to 95 degrees in Cinebench R23 , and held steady at that temperature with minimal variations at 93 and 94 degrees, tuning the turbo mode accordingly. That is the maximum temperature that AMD lists , so there is no doubt that this chip is reaching the limit with all its cores and threads active. However, in lighter loads the values ​​are totally safe as we can see in the attached graph.

As expected, consumption has also risen. We are not surprised, and the values ​​that we have registered in our tests are very positive if we take into account the performance improvement that AMD has achieved with the Ryzen 7000. In other words, the Ryzen 9 7900X consumes more than the Ryzen 9 5950X, but it also offers much better performance in single thread, and very similar in multithread.

Ryzen 5 7600X: Synthetic Test Performance

The Ryzen 5 7600X has the same IPC as the Ryzen 9 7900X, always taking into account the peculiarity that it has less L2 and L3 cache. Due to its lower number of cores and threads, it offers lower performance in applications that mainly depend on multithreading. It is totally normal high since we are talking about a processor that has half the number of cores and threads than that one.

As we will see, the Ryzen 5 7600X is an excellent option for those who want to use their PC mainly for gaming, although this does not mean that it is not capable of offering good performance in professional applications. We’re going to dig into this with the same round of testing that we’ve used with the Ryzen 9 7900X.

In CPU-Z we see that the Ryzen 5 7600X more than doubles the single-thread performance of the Ryzen 5 1600X , a chip that hit the market in 2017. This means that with Zen 4 AMD has managed to double the performance in just five years. The values ​​​​it obtains in this test are very good, although it falls a little below the Ryzen 9 7900X in single thread because its turbo mode has a slightly lower peak.

We now jump to see the performance in Cinebench R23 and we have a result of 1,924 points in single thread and 14,811 points in multithread . This leaves us only slightly behind the Ryzen 9 7900X’s 1,997 points single-threaded, and well above the Ryzen 5 5600X’s 1,578 points. It also overwhelms the latter in multithreading. With this test alone, the generational leap marked by the Ryzen 5 7600X compared to the Ryzen 5 5600X is clear, although in multithreading it loses against the Intel Core i5-12600K, which achieves 16,906 points.

Making a comparison between the Ryzen 5 7600X and the Core i5-12600K we see that the former clearly outperforms it in single thread with more than 7% difference, but the latter wins in multithread by more than 12%. This has an explanation, and that is that the Intel chip has a configuration of 6 high-performance cores + 4 high-efficiency cores, which translates into a total of 10 cores and 16 threads.

In PassMark we have achieved a score of 29,791 points , which places the Ryzen 5 7600X as one of the most powerful CPUs in the world. It is a very positive result, since it manages to slightly exceed the 29,320 points that we obtained at the time in this test with the Intel Core i5-12600K.

In V-Ray the Ryzen 5 7600X has achieved 11,382 samples , a score that far exceeds the 8,065 samples achieved by the Ryzen 5 5600X. The difference is once again substantial, and highlights that Zen 4 is a huge leap from Zen 3, even though we are only talking about a generation difference between the two architectures.

We are now going to see the AIDA 64 AES FPU performance test, which is expressed in MB/s. The Ryzen 5 7600X has achieved 147,843 MB/s, a score that is almost half of that obtained by the Ryzen 9 7900X, and which allows it to easily surpass the Ryzen 9 3950X. This means that it is an excellent result , without a doubt.

The AIDA 64 memory and cache test leaves us with positive results if we compare it with the data we have from the Ryzen 5 5600X. The caches are better tuned and thanks to the use of DDR5 memory we have a much higher bandwidth . There is nothing negative in the test, although as it happened with the Ryzen 9 7900X we received a notice that this version of AIDA 64 was not optimized to work with our team’s CPU, which is normal considering that it was not yet available when we made tests.

We now have to take a look at the Blender results and in general there are no surprises, the Ryzen 5 7600X registers scores that are equivalent to half of the values ​​obtained by the Ryzen 9 7900X. Yes, this is a positive thing, there is no doubt about that. I remind you that these scores refer to samples per minute, and not to rendering time.

We now turn to see the results that we have obtained in 3DMark CPU, and we have some data that fit perfectly when we compare them with those that we have obtained with the Ryzen 9 7900X. The difference between the two is minimal when we work with one, two and four threads , although the latter is a little more powerful in those three tests. When we get to the eight-thread test, the differences are already accentuated, and the Ryzen 9 7900X manages to impose itself convincingly thanks to its 12-core and 24-thread configuration.

In any case, the performance data we have is very good for a 6-core, 12-thread chip . The working temperatures are high when the Ryzen 5 7600X works at full power , although they do not reach worrying levels, and as expected, the working frequencies scale well depending on the number of active threads.

• With an active thread, the frequency reaches 5,444 MHz and the temperature stabilizes in the range of 58.24 degrees.
• With two active threadswe have a fairly stable frequency of 5,446 MHz and a temperature of 62.75 degrees.
• With four active threads, the temperature rises to 74.63 degrees and the frequency remains at 5,445 MHz.
• With eight active threads, the frequency remains at 5,446 MHz and the temperature rises to 79.50 degrees.
• With all the threads active, the temperature peaks at 82.38 degrees and the speed drops to 5,320 MHz.

We can enlarge the gallery below by clicking on it.

Ryzen 5 7600X: Gaming Performance

The Ryzen 5 7600X has set the bar very high in synthetic performance tests. There is no doubt that it is a very powerful processor, so much so that it is capable of giving a “pulse” to other chips with a higher core count thanks to its very high IPC, but how does it behave in games? I know you’re looking forward to finding out, and I’m not going to keep you waiting.

As I told you at the time when talking about the Ryzen 9 7900X, I have used four different resolutions, 720p, 1080p, 1440p and 2160p so that you have a clearer and broader vision of the performance that this processor is capable of offering and how the resolution does not affect only to performance, but also to CPU scaling. The lower the resolution, the greater the impact on the CPU , and vice versa.

We start with Red Dead Redemption and see that in 720p the Ryzen 5 7600X manages to beat the Ryzen 9 7900X by 1 FPS . In 1080p it achieves the same score as that one, in 1440p it exceeds it by 1 FPS and in 2160p both processors register 68 FPS on average. The Ryzen 5 7600X positions a bit better than the Ryzen 9 7900X in this game, clearly outperforming the Ryzen 9 5950X in 720, 1080p and 1440p.

We now turn to Shadow of the Tomb Raider. This game also has a strong dependency on IPC, but also scales well on multi-threaded processors, which helps us understand how the Ryzen 9 7900X can beat the Ryzen 5 7600X by so much at low resolutions. In 720p the Ryzen 5 7600X achieves 253 FPS on average, 9 FPS less than its older brother , and in 1080p it is 222 FPS, 6 FPS less than the Ryzen 9 7900X. At 1440p the difference is reduced to only 2 FPS less than the Ryzen 9 7900X, and at 2160p there is practically a tie.

Despite the differences that we see, the results obtained by the Ryzen 5 7600X are simply fantastic, and allow it to mark a very clear generational leap compared to the Ryzen 9 5950X, which, as we said before, achieves 196 FPS in 1080p, that is, the Ryzen 5 7600X yields 13.76% more.

Cyberpunk 2077 is another title with a strong CPU dependency, which thanks to the improvements it has received (patch 1.5) now scales better on CPUs with a high core count , although it is also very dependent on the IPC, something that we have been able to confirm in numerous tests, so keep that in mind.

At 720p the Ryzen 5 7600X achieves 178 FPS, only 5 FPS less than the Ryzen 9 7900X. In 1080p the difference is more marked with 136 FPS in the first compared to 152 FPS in the second, 16 FPS difference . The distances get shorter as we go up the resolution, but even at 1440p they remain considerably high as we have 8 FPS difference in favor of the Ryzen 9 7900X. In any case, it is a very positive result for the Ryzen 5 7600X, which outperforms the Ryzen 9 5950X in all resolutions. To illustrate this, we just have to remember that the Ryzen 5 7600X achieves 97 FPS in 1440p while the Ryzen 9 5950X achieves 89 FPS in that resolution, which leaves us with a difference of 8.99% in favor of the first.

In Death Stranding we have a very curious but very easy to explain scenario. We already saw that with the Ryzen 95950X we had a huge bottleneck even at 1440p , and this is again clear with the average frames per second recorded by the Ryzen 5 7600X.

As with the Ryzen 9 7900X there is no big upscaling until we go from 1440p to 2160p , but at least there are small differences in performance when going up to 1080p and 1440p, which is not the case with the Ryzen 9 5950X. Compared to the Ryzen 9 7900X, the Ryzen 5 7600X falls a bit behind , probably because the former scales its frequencies better in this title.

In Gears 5, the scaling achieved by the Ryzen 5 7600X is excellent, and it is practically as good as what we saw with the Ryzen 9 7900X . In general, the differences between the two are minimal, since the second achieves 3 FPS, in 1080p it wins by 1 FPS and, curiously, in 1440p it loses by 1 FPS compared to the Ryzen 5 7600X. That said, an excellent result from the Ryzen 5 7600X.

We finished gaming benchmarks with Metro Exodus Enhanced Edition. I have kept the same approach and settings as with the Ryzen 9 7900X, so we will be able to more easily make a direct comparison between it and the Ryzen 9 5950X. In this case, the results are curious, since the Ryzen 9 7900X performs worse in 720p and 1080p, but performs better in 1440p and 2160p, although in those resolutions the difference in its favor is only 4 FPS and 1 FPS. The advantage of the Ryzen 5 7600X over the Ryzen 9 5950X in this game is therefore very large.

We are going with a summary of relative performance to facilitate the comparison between the Ryzen 5 7600X and the Ryzen 9 5950X in games. As we can see, there is a difference of more than 12% in 1080p in favor of the first one, and this is reduced to almost 5% in 1440p . Not bad at all, especially considering that at that resolution the impact of the CPU on performance is considerably reduced.

Ryzen 5 7600X: Temperatures and consumption

With the Ryzen 5 7600X, a considerable rise in working temperatures compared to the previous generation is also evident. AMD recommends a mid-range air cooling system, but seeing the maximum temperatures that the sample we have used has reached, I would recommend accompanying it with a 240 or 280 mm liquid cooler.

The highest temperature value reached by the Ryzen 5 7600X occurred in Cinebench R23, where it remained at a stable average of 89 degrees . It is a high value, but safe. In the rest of the tests, the temperatures recorded were much better because in the end the total load was lower and it did not reach 100% CPU usage at any time.

Regarding consumption, we have a maximum average of 106 watts , also recorded in Cinebench R23. As expected, the rest of the values ​​remain at lower levels due to the lower workload. There is also a notable difference compared to the Ryzen 5 5600X, but in exchange for this higher consumption, we will enjoy a much higher performance.

Overclock Considerations

We have used an X670E platform, which means we can overclock the Ryzen 9 7900X and Ryzen 5 7600X . However, seeing the maximum temperatures that both reach when they reach 100% use, 95 and 89 degrees respectively, and considering how well the turbo mode scales in this generation, I consider that it has no value for most of the users, since we will end up getting into temperatures that can end up causing irreversible damage.

Thanks to the turbo mode, the Ryzen 9 7900X moves between 5.3 GHz and 5.4 GHz depending on the active cores and threads, and it approaches 5.7 GHz with one or two active threads . This dynamic adjustment offers a more balanced user experience in relation to performance and temperatures than forcing an overclock that in the end can compromise the integrity of the CPU. We have a very limited margin.

The Ryzen 5 7600X has a bit more overclocking headroom at the temperature level, but this is still unnecessary thanks to excellent turbo mode scaling, which seamlessly tunes to almost 5.45 GHz with one or two active threads , and stays on 5.3 GHz with all active threads without compromising consumption or frequencies. The overclock margin is also quite limited.

AMD has done an excellent job fine-tuning the working frequencies of the Ryzen 7000 , and has dared to reach such a high level that the overclock has lost almost all its meaning, something interesting since, as our regular readers will remember, this already had little value with the Ryzen 3000 “X” series, and even less with the Ryzen 5000 “X”, in fact when overclocking these it was common to lose performance in loads that did not use a large number of cores and threads because the turbo mode scaled a lot better.

Final Notes: Zen 4 has met our expectations

Since the debut of Zen architecture in 2017, the path that the Sunnyvale giant has maintained has been simply impressive. AMD has been able to significantly increase the performance of its processors with each new architecture, both single-threaded and multithreaded, and little by little has been polishing its MCM design until it has shown that it was not only perfectly viable as an alternative to the design of monolithic core, but also can overcome it.

Ryzen 5000 was the consecration of the MCM design, not in vain it is an architecture that finally allowed AMD to surpass the Intel Core Gen11 on all fronts, including single-thread performance, which was its only pending account. With the Ryzen 7000 series, this design has matured deeply, it has done so both at the IPC level and at working frequencies, and has allowed AMD to remove “a thorn” that it had pending , that of finally exceeding 5 GHz frequency.

In terms of raw performance, both the Ryzen 9 7900X and its little brother, the Ryzen 5 7600X, represent an important generational leap and are worthy heirs to the Ryzen 5000 legacy. That performance boost, and the leap to a next-generation platform compatible with DDR5 and PCIe Gen5, they are two very interesting keys that perfectly define the value offered by the Ryzen 7000 compared to the previous generation, but we must not forget that they are also the first to include an integrated GPU, and that they are accompanied AVX512 Instruction Manual.

With the Ryzen 7000 and the Zen 4 architecture in AMD they have managed to recover the crown of single-thread performance , which until now was in the hands of the Intel Core Gen12, and this has also allowed them to rise as the most powerful multi-thread generation . I have not been able to test the Ryzen 9 7950X, but seeing how well the Ryzen 9 7900X positions itself, and considering that it has four cores and eight threads less than that, it is something that does not support discussion.

It is true that temperatures and consumption have risen despite the reduction in the process, which is now 5nm, but we must bear in mind that this has been compensated with a significant improvement in performance, and with a huge increase in working frequencies. We cannot expect increases of between 800 MHz and 1 GHz to be totally free in terms of consumption and working temperatures.

As far as the platform and BIOS are concerned, I have to assess the health and performance of the version that will reach consumers , and not the early BIOS that we received in the test unit, since this will not fall into the hands of consumers. Yes, I had to mess around a bit to update to a more advanced build that would be the definitive one, but the experience with this one was simply perfect, the AMD EXPO profile loaded at the first click and the system stability was total.

The Ryzen 7000 are already available in Spain, and these are the official prices that AMD has confirmed to us:

• Ryzen 5 7600X from 369 euros.
• Ryzen 7 7700X from 489 euros.
• Ryzen 9 7900X from 669 euros.
• Ryzen 9 7950X from 849 euros.