PC of the month (November 2022)

The autumn months are traditionally full of interesting announcements and are not accidentally chosen by component manufacturers for the most extensive updates. The beginning of the active season is the best time for new decisions and gradual preparation for the hottest holiday period. In the next “PC of the month” we try to understand the current situation in the component segment and choose the best options game platform taking into account current needs and opportunities.

The article was prepared with the support of the TELEMART.UA store.
Assemble your “PC of the month” with handy configurator.

Basic gaming PC of the month

Ready configuration of the basic gaming PC in TELEMART.UA on Intel

Ready configuration of the basic gaming PC in TELEMART.UA on AMD

Significant events with the update of desktop platforms have not yet affected the basic configurations. No chips Intel Core of the 13th generation, nor the corresponding Ryzen 7000 processors under Socket AM5 for starting game assemblies. However, there is absolutely no cause for sadness here, because the available CPUs are quite capable of their tasks.

For Intel-based systems, we continue to use Core i3-12100F (4/8; 3.3/4.3 GHz). Yes, indeed, a 4-core chip as part of a gaming platform at the end of 2022 looks a bit illogical, but numerous practical tests show that in connection with a graphics card, even the average level of performance of the basic Alder Lake is sufficient. Not a lot of stock, but it gets the job done, especially considering the $115 price tag. In addition, the chip is extremely economical, which is especially relevant nowadays, and also allows you to at least try to use a processor with a standard cooler without additional costs.

As for the migration to the 13th generation Core chips, it remains a prospect for PCs of this class. In addition, unfortunately, it will not bring special bonuses. As expected, for younger chips of the Core i3/i5-13xxx family, crystals with the Alder Lake architecture will actually be used. And if in the case of Core i5, additional energy-efficient cores will be unlocked for all models, then Core i3 will remain with the functional formula 4/8 and will receive only a small increase in operating frequencies. No holiday. Therefore, initially focusing on Core i3, you can take Core i3-12100F without regret, without waiting for Core i3-13100F to be officially presented. An additional 100-200 MHz will have almost no effect on overall performance.

If the platform is used for work where maximum multi-threaded performance is important, the 6-core Core i5-11400F for $155 and Core i5-10400F for $135 remain relevant. But in this case, you will have to stay on the LGA1200 platform and limit yourself to Core 11th generation chips during possible future upgrades.

In the case of the Core i3-12100F and, accordingly, the LGA1700 platform, we recommend choosing from boards based on the Intel B660 chipset. For $110, you can pick up a model with moderate equipment, quite suitable for a platform of this class.

If you’re trying to save money on a motherboard, you can find Intel H610 models for as low as $75. But often even a visual inspection of such devices is enough to give preference to more equipped versions. Savings are felt on everything, starting from primitive circuitry and layout of power units, ending with limited functionality and the simplest cooling systems of the chipset. Yes, equipping relatively inexpensive boards with Intel B660 is also far from an example to follow, but still it is better there. VRMs sometimes already have radiator blocks, and the ability to use memory modules in modes higher than DDR4-3200 already justifies the additional costs in itself.

For a basic gaming platform on AMD, there are two excellent options that we are quite satisfied with – the Ryzen 5 5500 (6/12; 3.6/4.2 GHz, 16 MB L3) and the Ryzen 5 3600(6/12; 3.6/4.2GHz, 32MB L3). Although both 6-core chips have fundamentally different internal structure and layout, in the end they offer close performance indicators, both in games and in applications. Therefore, we give a minimal preference to the more modern and economical Ryzen 5 5500. By the way, over the last month, the price of this model has decreased by $15 to $110. It seems like a small thing, but a nice touch when building an entry-level gaming system on a very limited budget.

As for the Ryzen 5 3600, this chip should also be given credit, it is no coincidence that it still holds an honorable place in the current popularity ratings of processors. Belonging to the Matisse line with Zen 2 architecture will not be a disadvantage here, given the 32 MB of L3 cache, PCI-E 4.0 support and especially the price of $110.

Motherboards for the Socket AM4 platform with similar equipment usually cost less than models for LGA1700. Although some segmentation needs to be taken into account here. For example, simple models on the AMD B550 could be purchased for a long time for $20-40 cheaper than devices on the Intel B660. However, over the past few months, the price of Ryzen boards has increased slightly, and the price delta has decreased to $10-20. In general, it will take $90–100 to purchase a suitable model for AMD B550. In general, you can find boards on the AMD B450 chipset on sale for $60-65, and even with official support for the 16-core Ryzen 9 5950X. But – no, thank you.

Choosing a graphics card for basic configurations is a special act, especially after the hungry period of the Miner turmoil. On the one hand, you want to get a model with the best price/performance ratio, on the other hand, it is also important to ensure a certain margin of speed to be able to get comfortable fps even in difficult projects. Therefore, in view of the perspective, sometimes it is more rational to raise the conditional price limits a little.

Last time we complained that there are almost no offers for GeForce GTX 1660 6 GB models on the market, and those that are available are too expensive. Although our price wishes of $230-240 have not yet been fully met, the cost of the basic versions has really come down significantly. Although the range of devices in this line is small, some models can already be purchased for $250-260. In this case, this is a working option for the basic configuration. Of course, you always want to get more fps and have a better margin of performance, especially in this class of systems, but financial opportunities are often quite limited at this time, so it is quite possible to pay attention to the GeForce GTX 1660 6 GB, it will work for the indicated price.

However, if there is an opportunity to add $20, it is better to do so by focusing on the GeForce GTX 1660 SUPER 6 GB. Basic versions of the line can be found for $270. We will remind you how SUPER-versions differ from basic modifications. The difference is only in the type of memory chips. If the starting version is equipped with GDDR5 with an effective frequency of 8,000 MHz and a total bandwidth of 192 GB/s, then the SUPER modifications received GDDR6 with an effective frequency of 14,000 MHz. In this case, the memory bandwidth increased up to 336 GB/s. The rest of the specs remained the same, but just the increase in GPU turned out to be enough to improve gaming performance by 10-15%. Therefore, a certain extra payment for the GeForce GTX 1660 SUPER version is quite justified.

The selection of available GeForce GTX 1660 Ti 6 GB models is much more modest than SUPER modifications. Taking into account the fact that the heart-breaking “commemorative” spurt of the latter allowed them to come almost close to the Ti-versions, it makes little sense to pay an additional $30-40 for the “titanium” version.

If you have already decided to significantly speed up the graphics subsystem in the basic configuration, then Radeon RX 6600 8 GB can be a reference in this case. In recent months, these adapters have significantly decreased in price. Select models are already available starting at $320. Yes, the difference in price is also quite large, but this is really a case when the end justifies the means. On average, the Radeon RX 6600 will be 30-40% faster than the GeForce GTX 1660 SUPER/Ti. In addition, the model allows you to play with the activation of ray tracing. For this class of systems, this can hardly be called a “killer feature”, but a significant acceleration of traditional rendering will definitely be felt without direct performance measurements. However, $320–330 is already a considerable amount for a graphics card for the starting game configuration.

Allocating such funds for the purchase of a graphics card, you can also look at the GeForce RTX 3050 8 GB. But in this case, our sympathies are with AMD’s solution. The base model of the GeForce RTX lineup has only a minimal advantage over the GeForce GTX 1660 SUPER/Ti when using traditional rendering methods. The presence of hardware RT-blocks and tensor cores really expands the possibilities and allows you to experiment with tracing and DLSS in cases where they are implemented by developers. However, the power of graphics cards of this class is limited, so raw graphics power remains a priority.

Savings options in the form of a GeForce GTX 1650 4GB and a Radeon RX 6500 XT 4GB for $180-200 would not be considered in that case, let alone a GeForce GTX 1630 or Radeon RX 6400. These were anti-crisis forced solutions in case of an emergency. The mining euphoria is over, so now it is better to choose from those models that will really attract even difficult projects, even if not with the maximum quality of graphics. However, of course, now there are also enough reasons for greater savings.

As an option, if the budget is significantly limited, you can take a closer look at graphics cards with factory recovery. Usually, such models have a limited warranty period (3-6 months) and may be released in unpresentable packaging, but the price can be very interesting compared to the same for a new model.

A risky, but even more financially attractive option is the secondary market. Now aggregators are simply flooded with private ads for the sale of graphics cards, which were mostly used for mining electronic coins. Everything related to the purchase of such devices is a lottery, according to the results of which you can either get a fully working video card without any issues with stability, or spend your money in vain.

Optimal gaming PC of the month

Intel Core i5-12400F (6/12; 2.5/4.4 GHz)$200

Ready configuration of the optimal gaming PC in TELEMART.UA on Intel

Ready configuration of the optimal gaming PC in TELEMART.UA on AMD

In optimal configurations, there is also complete idyll and certainty with processors. One can only complain about the unpleasant price increase of Core i5-12400F (6/12; 2.5/4.4 GHz), which still occupies a place in systems of this class. The hexa-core chip has added about $15 over the past month, so purchasing a popular model will cost ~$200.

What are the prospects for strengthening this position in the near future? In theory, this is a transition to the 13th generation Core. As expected, Intel presented and started selling chips of a new family of enthusiast models, while the appearance of regular versions does not even have a set date yet. They may appear before the end of this year, but it is also possible that the manufacturer will delay their launch until January CES 2023.

As for the immediate successors, the place in the optimal configuration will most likely be taken by Core i5-13400F. Although Intel did not make public the future full range of chips, partners did it “inadvertently” for them. However, there are still some questions about the technical characteristics. Likelyfor chips from Core i5-13600 and younger, crystals with the Golden Cove (Alder Lake) architecture will be used, and not the updated Raptor Cove. It’s a bit of a shame, of course, but despite this, the new Core i5 will still get a certain acceleration, and not only thanks to the increased frequencies.

Energy-efficient cores (E-Cores) will be unlocked for all 6-cores. The Core i5-13400/13400F modifications will get 4 additional computers, while the older Core i5-13500/13600 models will get 8 Gracemont cores. The additional resources will definitely affect performance in multi-threaded applications. How useful this modernization will be in games, we will see after practical tests. At least the increase in the number of cores, albeit not so powerful, should close the issue with the potential lack of capabilities of 6-core CPUs in the most resource-intensive projects.

Thus, if the actual price of the Core i5-13400F will not be significantly different from that of the Core i5-12400F, then the 13th generation model will usually be more desirable in the optimal configuration. In addition, existing motherboards will be suitable for its use, so there are no extra costs here. Therefore, it is a matter of time – to wait for the official announcement and the start of sales of new chips.

By the way, as for the boards themselves, for an optimal system, we also recommend choosing from Intel B660 models. Moreover, in order to meet the class, more equipped devices with better functionality and a VRM unit with additional cooling are requested.

For an AMD-based system, we consider the Ryzen 5 5600 (6/12; 3.5/4.4 GHz) to be the best option. Some time after the announcement and release, this model retailed for only $10-15 less than the older version of the Ryzen 5 5600X (6/12; 3.7/4.6 GHz), so it and did not cause much interest. But recently Ryzen 5 5600 has dropped in price rapidly. You can already buy this CPU for $160. This is an extremely attractive price for a 6-core 12-thread Vermeer family chip with Zen 3 architecture and 32 MB of L3 cache. Numerical tests show a minimal difference in performance compared to the Ryzen 5 5600X. The decrease in performance is due to a lower operating frequency of only 100–200 MHz. Otherwise, it is in no way inferior to the X-version. But the price difference of $40–50 is quite noticeable even for mid-level systems. We see no reason not to save, here is a very good opportunity.

As for migration to Socket AM5, there is unfortunately nothing to please potential owners of optimal configurations. The cheapest processor of the new line – Ryzen 5 7600X (6/12; 4.7/5.3 GHz) – costs about $330 in Ukraine. Which is quite humane, given the recent launch and $300 MSRP. But this is too much for our configuration. In addition, new motherboards even with the AMD B650 chipset still cost $200+ and the selection is limited to only a few models. And we have not yet taken into account DDR5 memory, the purchase of which will also cost more than DDR4 of the same capacity. In summary, we can say that Ryzen 7000’s time for optimal configurations has not yet come. Here, the positions of Socket AM4 solutions are very powerful. A good price-performance ratio allows you to get a balanced platform for reasonable money. Therefore, we are watching the development and waiting for more attractive offers.

Now let’ discuss a little about video cards for optimal configurations. It must be admitted that the large range of available options is a bit spoiling and requires more time to be spent on choosing the best graphics adapter. Now the trivial departure “more expensive is better” generally works without any particular failures, but no one canceled the rational approach in the case of optimal configurations.

Last time, we offered to consider two models – GeForce RTX 3060 12 GB and Radeon RX 6650 XT 8 GB, which in general can be considered the backbone of this class of PC. Over the past month or a half, video cards of these lines have minimally lost in price ($10–20), so they have become even more attractive to purchase. We remind that the model on the AMD chip usually has an advantage of 10-15% when using classic rendering, while the adapter with the processor from NVIDIA will look better when playing with ray tracing. These are worthy opponents and completely acceptable win-win options, which you can safely choose, focusing only on your own preferences.

But this time we offer a closer look at another pair – GeForce RTX 3060 Ti 8 GB and Radeon RX 6700 XT 12 GB. The cost of more productive video cards continues to gradually decrease, so when choosing, there is a natural desire to get a faster model. In this case, moving to the next step in the graphic hierarchy will increase the number of fps by approximately 20-25%. Given that RTX 3060 Ti 8 GB and Radeon RX 6700 XT 12 GB can be found for $480-500, and the purchase of younger models will require ~$430, there is a good reason to consider all options. The performance/price ratio of older adapters is at least not worse, or even a little better, and this is a good argument.

The GeForce RTX 3060 Ti 8GB and Radeon RX 6700 XT 12GB are especially recommended if you’re going beyond Full HD and plan to play at 1440p. There you will definitely need additional speed.

As for the direct comparison of the capabilities of GeForce RTX 3060 Ti 8 GB and Radeon RX 6700 XT 12 GB, everything is quite interesting here. In classic rendering, they show almost identical results. When using RT, “green” will be more agile. Given this balance of power, this is already a somewhat atypical situation. On the side of the Radeon RX 6700 XT remains a larger capacity of local memory, which will not be superfluous at 2560×1440.

AMD usually takes such a situation into account, and therefore it is no coincidence that back in March, the developer presented an accelerated modification of the Radeon RX 6750 XT 12 GB. Due to a small GPU boost and an increase in the effective memory frequency from c 16,000 MHz to 18,000 MHz (+12.5%), it was possible to increase the overall performance by 5-10%. However, this model did not replace the previous one, but became another position in the line, and with a significantly higher recommended price at the time of launch ($479 vs. $549). Now the actual difference at retail is $30-40, the extra payment is justified unless you get a more efficient and quieter cooling system. However, of course, you should monitor the market, interesting offers are quite possible now.

Returning to the GeForce RTX 3060 Ti 8 GB, we can recall a certain negative precedent, which in the circles of enthusiasts, or rather miners, cast a shadow over the entire line. We are talking about isolated cases of problems with the operation of memory chips of one of the series of microcircuits manufactured by SK Hynix, which were manufactured at a certain time in 2021. After a long load, the chips did not always work stably even at standard frequencies. As it turned out, these were microcircuits from the test batch that went into production. However, even in this case, the problem did not acquire a mass character, but was “heated up” and received publicity. The probability of encountering something similar is vanishingly small, so this is definitely not a reason to abandon an interesting decision, fearing similar consequences. In addition, an additional reassuring factor is a 3-year warranty from graphics card manufacturers.

NVIDIA is trying to expand its range in the segment of mid-level graphics adapters. As expected, Californians without much fanfare announced the appearance of GeForce RTX 3060 8 GB and GeForce RTX 3060 Ti GDDR6X. As for the GeForce RTX 3060 8 GB, to reduce the memory capacity, the developers also needed to reduce the bus bit rate from 192 to 128 bits. This significantly affected the bandwidth, and therefore will obviously affect the overall performance of the graphics card. However, much will depend on the actual price of such models. If NVIDIA and partners can balance their offerings, why not.

In turn, the GeForce RTX 3060 Ti, on the contrary, received an accelerated modification. GDDR6X memory with an effective frequency of 19,000 MHz and a bandwidth of 608 GB/c instead of GDDR6 (14,000 MHz, 448 GB/c) will surely add performance to the graphics card. Quite a logical step by NVIDIA, which will allow to strengthen the competition with the same Radeon RX 6700/6750 XT.

NVIDIA did not specify the recommended prices for new modifications of video cards, noting that the corresponding models will go on sale in November. Therefore, we are waiting for the appearance of adapters in Ukraine, after which we will determine their prospects for inclusion in our recommendations.

By the way, the update of the GeForce RTX 3060/3060 Ti is a clear indication that NVIDIA is not planning an RTX 4000 series video card for this segment in the near future. Obviously, there are still plenty of GPUs with the Ampere architecture, so for a few more months it will be the adapters that will close the positions in the middle class on their basis. The same applies to AMD, which by the end of this year will definitely not offer its options based on RDNA 3 for optimal configurations.

But who else will compete for the attention of moderate players looking for balanced solutions, this is the Intel company, which has finally presented its video cards Intel ARC A770 8/16 GB and Intel ARC A750 8 GB. The models turned out to be quite interesting, although with certain features. The performance of adapters is quite difficult to compare, as it depends on several factors and specific projects. If you still try to average, the Intel ARC A770 mostly competes with the GeForce RTX 3060 12 GB. However, in some cases it can surpass even the GeForce RTX 3060 Ti, in others it is noticeably inferior to the younger model of the series. The ARC A750 has similar performance, only 5-10% behind the A770.

In general Intel adapters show better results when using DirectX 12, and in general there is a feeling that the chip is a “heavier” project, the higher the positions of new products compared to models from AMD/NVIDIA. In games with DX11 and earlier APIs, the performance of Intel graphics cards is more modest.

However, the performance of the A750/A770 when ray tracing was activated was quite pleasant, and the first feedback from the implementation of Intel XeSS intelligent scaling technology was quite positive. The software component is generally at a fairly decent level, especially for a start. However, it was not without certain roughnesses, and software options for setting/management are still inferior to software from opponents.

Intel ARC A750/A770 features also include the almost mandatory use of ResizableBAR technology for direct processor access to GPU memory. Therefore, Intel recommends using platforms with Core 10th generation chips or Ryzen 3000 processors and newer. It is precisely these systems that have ReBAR support. Without it, the performance of the video card can drop by a quarter or even more. Therefore, Intel ARC A750/A770 are initially not suitable for upgrading old platforms, while in new ones you need to remember to enable the corresponding option.

One of the features that the first reviewers also noted is the high power consumption of the video card in sleep mode. Adapters can consume 35-40 W without load, while devices from AMD/NVIDIA in such conditions do not need more than 10-15 W. To reduce such indicators, Intel recommends making certain settings in the BIOS and checking the power saving management schemes in the operating system. That is, of course, we could not do without the initial nuances.

Intel ARC A750/A770 with recommended prices of $289 and $329 are very interesting objects for experiments. Especially among enthusiasts who have been waiting for the appearance of fundamentally new video cards for a long time. And from this point of view, the novelties really live up to expectations, having extraordinary features. However, it is still too early to say that these devices will be optimal for the average consumer, who is primarily interested in the stable predicted performance of video cards.

In addition, with genuine interest, it is problematic to purchase Intel video cards even on Western online sites, not to mention Ukraine. And this is taking into account the fact that official sales have already started more than a month ago. We assume that the first batches of adapters have quite limited volumes. In addition, the manufacturer is trying to implement more reference models, while the number of partners is surprisingly very small.

Among video card manufacturers, we saw ARC A770/A750 devices from ASRock, and Acer, for which the production of such devices is not characteristic in principle. The Chinese Gunnir can also be included among the few manufacturers. We can only guess about the reasons for such modest initial support. Be that as it may, we follow the development of events with interest, and we are happy to test the ARC A770/A750 when we have the opportunity. Along the way, we wish Intel to withstand the initial pressure, find resources for optimizations and overcome skepticism about the success of such a necessary undertaking. A third player in the discrete graphics market will definitely not be superfluous.

Progressive gaming PC of the month

Ready configuration of the progressive gaming PC in TELEMART.UA on Intel

Ready configuration of the progressive gaming PC in TELEMART.UA on AMD

In the case of advanced gaming configurations, we are really pleased with the many excellent CPU options for platforms in this class. After the launch of Intel Core chips of the 13th generation, the range of current processors has expanded significantly. The priority in the conditional rating may change, but in our opinion, the Core i5-13600KF (6/12+8) is the most attractive for a powerful gaming system, but not for all the money in the world. Chips of the Raptor Lake family have significantly increased performance through internal optimizations and increased operating frequencies.

Despite the fact that the Core i5-13600KF has six powerful cores with Hyper-Threading and 8 additional economic cores, the top chip of the new Core i5 line, according to the results of many tests, is not inferior to the more expensive Core i7-12700KF ( 8/16+4), or even has a slight advantage (~5%) over the latter. This applies to both multi-threaded tasks and games. Core i5-13600KF at the start of sales is about $50 cheaper than the older chip of the previous generation – $340 vs. $390. Not that this is a decisive amount in the case of a progressive configuration, but it can partially compensate for the purchase of a cooling system, which will be needed in any case. It should be noted that the energy appetites of the Core i5-13600KF are not less than those of the Core i7-12700KF, so an effective cooler or SRO cannot be dispensed with here.

Will the power of a 6-core chip be enough for resource-intensive games? There are definitely no problems with actual projects. The thing is that you can no longer call the current Core i5 “6-core”. As we have already mentioned, the top model of the series has 8 additional energy-efficient E-Cores, which were also accelerated in the new CPU generation due to twice the capacity of the L2 cache memory and increased frequencies. Therefore, the Core i5-13600KF can simultaneously process up to 20 data streams.

If you need even more processing power for work tasks, or are worried about the potential lack of gaming performance of the “6-core” for future projects, then for such a case there is already a Core i7-13700KF (8/16+8) for $450. In multi-threaded tasks it will really improve results. A certain increase in fps can also be observed, but this is rather not due to an increase in the number of cores, but due to increased maximum frequencies by 1-2 cores – 5.4 GHz vs. 5.1 GHz, and the difference in L3 cache memory capacity is 30 MB vs. 24 MB. By the way, the power consumption of the Core i7-13700KF is even more noticeable than that of the Core i5-13600KF. Under a multi-threaded load, the indicators may differ by 60 W. If such computing power is really necessary, then it is not a question, but such a feature should be taken into account when choosing a cooling system.

I would like to get the new regular Core i7 models with a TDP of 65 W as soon as possible. For example, the same Core i7-13700F, which could become a compromise option both in terms of power and energy consumption, and in terms of price. But so far there are no release dates for Core 13th generation models for those who do not chase maximum megahertz at any cost.

Touching on the topic of operating frequencies, we cannot but mention a landmark event related specifically to the new 13th generation Core chips. Enthusiast Elmor managed to launch the flagship of the line – Core i9- 13900K at 8812.75 MHz, thereby breaking the world record. For a long time, flamboyant “dinosaurs” like the AMD FX 8370, which managed to boost from 4 GHz to 8700+ MHz, were at the top. It took more than 8 years to break these frequency bars, although history is silent on how many vats of liquid nitrogen were spent to set a new record. However, these are the toys for extremists, for whom such efforts are fully justified.

When choosing a motherboard for a progressive gaming configuration, we recommend considering Intel Z690 models. Devices on the Intel Z790 are also usually suitable, but in fact do not offer any innovations. Conditionally, the new generation of the top chipset only provides for the redistribution of the internal lines of the PCI Express bus in favor of PCI-E 4.0, and potentially allows the use of one more USB 3.2 Gen 2×2 port with a bandwidth of up to 20 Gb/c. Frankly, there is little improvement. Therefore, here it is rather necessary to take into account the general functionality of a specific board model, pay special attention to the capabilities of the CPU power stabilization subsystem. New powerful processors are quite a challenge for VRM. The choice between the Intel Z690 and Z790 is not fundamental, except that the models on the latter chipset will support the 13th generation Core by default, while the Intel Z690 boards will require a firmware update. However, the choice of models on the Intel Z690 is still much wider, but in general there are no objections to the Intel Z790.

The only fundamental decision before buying is DDR4 or DDR5. Core 13xxx processors support work with modules of both one standard and another, but motherboards work with either one or the other. DDR4 modules are still significantly cheaper. A set of 32GB (2x16GB) DDR4-3600 CL18 costs about $130, while a good set of the same capacity DDR5-6000 CL36 will cost already ~$250.

The launch of the Socket AM5 platform and the Ryzen 7000 chips based on the Zen 4 architecture certainly became the most significant event of recent times for the markets of components for desktop systems. A new processor socket, support for DDR5 and PCI-E 5.0, and a CPU with significantly increased clock performance. All this can be added to the asset of new AMD solutions. However, at the initial stage, the new Socket AM5 platform turns out to be quite a valuable pleasure, and even when it comes to assembling progressive configurations. The most affordable chip of the new line – Ryzen 5 7600X – having a recommended price of $299, appeared on sale for ~$340. Potentially interesting for a progressive configuration, the 8-core Ryzen 7 7700X retails for about $450 in Ukraine.

Motherboards with the AMD B650 chipset start at ~$200, and at least some variety of choice appears when you spend $250. More interesting in terms of equipment, models with AMD X670 already cost $300+, well, the most functional models with AMD X670E are available for maximalists for $500+. In addition, the kit will definitely need DDR5, and as we already mentioned, a good set of DDR5-6000 memory with a capacity of 32 GB, optimal for new AMD chips, costs about $250. In this case, the overall cost estimate for AMD’s new platform is somewhat disappointing. Especially in the presence of powerful solutions of a competitor who managed to somewhat spoil the celebration of the appearance of Socket AM5. In addition, it seems that AMD itself somewhat complicates the launch of the new platform by lowering the prices of chips for Socket AM4. Although, of course, this is a forced measure, since the developer needs to respond to the aggressive actions of the opponent.

If you look at CPU popularity ratings, it is Socket AM4 processors that attract the attention of many users. A great combination of performance/cost has a special value, regardless of the category of systems. We are sure that in the future solutions for Socket AM5 will be used for progressive configurations, but for now we prefer Socket AM4, especially since after the price reduction there is a particularly attractive option for a powerful gaming system. Of course, we are talking about Ryzen 7 5800X3D (8/16; 3.4/4.5 GHz; 96 MB L3). After reducing the retail price from $450 to $350, the technologically unique 8-core model with a giant L3 buffer has become even more interesting in the context of a gaming platform.

Colleagues from the TechPowerUp resource recently conducted a large-scale test with 53 games to compare the performance of Ryzen 7 5800X3D and Ryzen 7 5800X. According to the results, the model with 3D V-Cache had an average advantage of 19% when using the Full HD mode. Moreover, in most projects the difference is even more significant – in some cases up to 50%, quite often +20–35%. To get that kind of acceleration within a single platform is incredible. In addition, the Ryzen 7 5800X3D, despite the capacity of the L3 cache, is a fairly economical chip that does not require a board with an extremely powerful VRM or an expensive supercooler. Its power consumption is lower than the base model Ryzen 7 5800X.

Together with the Ryzen 7 5800X3D, it is quite possible to use motherboards on the AMD B550 or relatively inexpensive models on the AMD X570/X570S for $150-170. In the latter case, it is better to choose from options without active cooling of the chipset.

As for the video card, there are no changes here. GeForce RTX 3080 10 GB and Radeon RX 6800 XT 16 GB remain working options even despite the loud announcements. From the line on the NVIDIA chip, you can choose a suitable model for $820-850. GPU adapters from AMD are slightly more affordable – $750-800, and some promotional offers allow you to get such models even for $670-700.

At the end of September, they announced GeForce video card RTX 4090 24 GB, and on October 12, models based on the top GPU AD102 with Ada Lovelace architecture have already gone on sale with a recommended base price of $1,600. The ultimate graphics adapter that has raised the bar significantly, especially in 4K mode. The increase in gaming performance is about 50-60% compared to the flagship of the previous generation line – GeForce RTX 3090 Ti 24 GB. Formally, both video cards rather belong to computing devices, but recently even high prices do not deter the most demanding gamers, especially those who managed to purchase a 4K monitor. Increased interest in such graphic monsters is confirmed at least by the fact that since the announcement, NVIDIA has already shipped more than 100,000 AD102 chips to video card manufacturers.

With the release of the GeForce RTX 4090, gamers can finally get over 60fps in 4K modes with tracing enabled, even in cases where projects don’t have an intelligent scaling option. By the way, GeForce RTX 4xxx models received support for the new generation of DLSS 3.0 with AI algorithms for generating additional frames, which promises an even greater increase in performance.

GeForce RTX 4090 is a fairly energy-intensive video card. For adapters, a TDP of 450 W is declared, but in practice, the peak consumption of models forced by manufacturers can be up to 550 W. All models based on the AD102 are equipped with a new 12+4-pin 12VHPWR connector for connecting additional power. According to the standard, the connector is designed to supply 600 W. But rather compact dimensions, and especially the nuances of implementing adapters from four 8-pin connectors sometimes lead to unpleasant consequences. More than a dozen Reddit visitors have already encountered the problem of overheating and melting of the contact group. All cases are recognized under warranty, NVIDIA collects statistics and, together with video card manufacturers, tries to find out the causes in more detail and eliminate them in the future. Such are the problems of players with a tight wallet.

GeForce RTX 4090 video cards are already available in Ukraine, but the first price tags are unlikely to appeal even to those who are ready to buy an expensive “toy”. Unfortunately, the reference models are not available here, but the versions from partners start at the equivalent of $2,200, while the top forced modifications ask for about $3,100. It is surprising, but it seems that the mining boom is behind us, and the desire to skim the cream and get “two meters” remains. If you want to count how many game consoles you can buy for the price of one video card, please comment. And the total amount for the purchase of a progressive configuration is more modest here. However, here everyone chooses their own. The desire to own the best can turn off the rationality, but we will definitely not dissuade you. The main thing is to pay for the consumed electricity on time (when it is available), and not to excessively bend the power cableale near the connector.

For enthusiasts with slightly smaller demands, NVIDIA introduced the GeForce RTX 4080 16 GB. The model is already based on the AD103 graphics processor, which includes 9728 computing units, so it is equipped with 16 GB of GDDR6X memory, working at an effective 22,400 MHz. With a 256-bit bus, the bandwidth is at the level of 717 GB/c. The recommended price of the GeForce RTX 4080 16 GB is $1,200. According to the test results, this model is 25-30% inferior to the GeForce RTX 4090 in 4K mode, but about 50% faster than the GeForce RTX 3080 and ~15% ahead of even the RTX 3090 Ti. Well, it looks quite promising, although one cannot help but notice a significant increase not only in performance, but also in the cost of new generation video cards. We will remind that the starting recommended price of GeForce RTX 3080 10 GB in 2020 was $700.

During the announcement of the new generation video cards, the list of devices also included the GeForce RTX 4080 12 GB model with a price tag of $900. NVIDIA planned to use the AD104 chip for this version with a significantly reduced number of computers, bus bitrate and memory bandwidth. Such intentions caused considerable resonance, because in this case video cards of the same series would differ significantly in performance (20-30%). To their credit, the developers took the criticism and eventually canceled the launch of the GeForce RTX 4080 12GB version. Moreover, NVIDIA even compensated manufacturers for the costs of preparing for the release of this model. However, instead of canceling, it would certainly be better to see a similar model, for example, under the designation GeForce RTX 4070 Ti with a slightly lower recommended price tag. Still, we need more Ada…Lovelace. But the Californians decided to hold an “unlaunching” while not offering anything to replace it. Perhaps the activities of the main competitor in this segment will accelerate the necessary processes.

According to the latest leaks, the canceled video card will still appear under the label GeForce RTX 4070 Ti 12 GB. Well, this is a logical move, but the source claims that such models will not be introduced until January next year. Obviously, this will happen simultaneously with the appearance of the basic GeForce RTX 4070 model, which will also be based on the AD104 graphics processor in a lighter version.

On November 3, AMD finally held a long-awaited event, presenting video cards Radeon RX 7900 XTX 24 GB and Radeon RX 7900 XT 20 GB based on the Navi 31 chip with RDNA 3 architecture. The graphics processor uses a chiplet layout with the main crystal (5 nm, TSMC) containing the computing part and six additional Memory Cache Dies (MCD) chiplets with memory controllers and second-generation Infinity Cache buffer with a total capacity of 96 MB. In the case of the Radeon RX 7900 XTX, the Navi 31 processor includes 12,288 computers and operates at a frequency of up to 2.5 GHz. At the same time, the maximum theoretical computing power is 61 TFLOPS (single precision). A 384-bit bus and GDDR6 at an effective 20,000 MHz provide a bandwidth of 960 GB/s. The video card has a TBP of 355 W and an MSRP of $1000.

The younger model Radeon RX 7900 XT 20 GB uses the same processor, but already with 10,752 active computers and one deactivated MCD chiplet. In this case, the bit rate of the bus is already 320 bits, and the memory bandwidth is 800 GB/s. It is interesting that AMD uses GDDR6 for the top adapters, and not the even faster GDDR6X. However, as a result of the wide tire, it is possible to obtain quite good final indicators of PZP. The power consumption of Radeon RX 7900 XT should be up to 300 W, while the recommended price of this model is $900.

The relatively small price delta between the Radeon RX 7900 XTX 24 GB and the Radeon RX 7900 XT 20 GB, which is probably less than the actual difference in their performance, rather immediately directs potential buyers to the older model. As for the capabilities of top video cards in general, AMD did not compare them with the performance of competitor graphics adapters in its presentation. The company indicates that the Radeon RX 7900 XTX in 4K mode will be 50-70% more productive than the Radeon RX 6950 XT. It will obviously be a competitor to the GeForce RTX 4080 16GB, while there are no alternatives to the GeForce RTX 4090 yet.

Of the interesting points, it should be noted that the new AMD video cards use 8-pin auxiliary power connectors. Reference models have two such connectors, while original forced adapters from partners can probably be equipped with even three. Even before the official announcement, it will be known that the developers do not plan to use 12VHPWR for Radeon 7000.

Another interesting fact is that the Radeon RX 7900 XT/XTX, as well as new NVIDIA video cards, use PCI-E x16 4.0. And this means that for the next 1.5-2 years PCI-E 5.0 nand motherboards will not be needed in full.

But what pleased the new AMD video cards was the implementation of DisplayPort 2.1, which allows you to display a picture with a frequency of up to 480 Hz in 4K mode, and up to 165 Hz in 8K. While NVIDIA’s top models only support DisplayPort 1.4a with much lower bandwidth. Although the same GeForce RTX 4090, faster ports would not be superfluous at all, since in some projects it can already provide more than 200 fps in 4K mode.

Coming back to the Radeon RX 7900 XT/XTX, we eagerly await the results of independent testing. Video cards will have to wait a little longer – the start of sales is scheduled for December 13. The models look quite attractive and interesting from a technological point of view. In addition, AMD promises to implement FSR 3 in 2023 against NVIDIA’s DLSS 3.0, so the competition should intensify on the graphics AI front as well.

If in September “PC of the month” we had quite a lot of unknown and important expected announcements, this time there is much more certainty and opportunities to consciously choose certain components for assembling powerful game systems. Even basic configurations are well prepared for challenges, while optimal and progressive PCs will allow you to get more comfort. Collecting expensive systems at this time, you should not forget about the support of the Armed Forces, this will speed up our victory.

The article was prepared with the support of the store TELEMART.UA.

Assemble your “PC of the month” with handy configurator, which allows you to quickly choose a PC, get advice, choose one of the powerful EVOLVE computers, or create your own configuration.