AMD RDNA 3 details: architecture changes, AI acceleration, DP 2.1

November 3 saw the expected unveiling of a new generation of Radeon graphics cards from AMD with RDNA 3 architecture, a competitor to the GeForce RTX 4000, the already selling new GPU generation from Nvidia. In this article, we’ll discuss the chiplet design of the new 5nm AMD GPUs, the changes in architecture, and the new functionality and features that they bring. RDNA 3 has a lot of new stuff compared to the previous generation. Read more “AMD RDNA 3 details: architecture changes, AI acceleration, DP 2.1” »

65W CPU with Zen 4 architecture is coming: AMD Ryzen 7 7700

When the specs of the Ryzen 7000s first leaked, one thing was disappointing. These processors use 5nm process that is much more efficient, yet AMD has not exploited this potential. Likely due to pressure from Intel’s big.LITTLE processors, the company has increased power draw to raise multi-threaded performance and instead of 65W and 105W TDPs, the TDPs went up to 105W and 170W. But lower-power versions are fortunately in the pipeline. Read more “65W CPU with Zen 4 architecture is coming: AMD Ryzen 7 7700” »

Why is Zen 4 so fast in Topaz Labs AI apps? In fact it’s Intel’s doing

Ryzen 7000 with Zen 4 architecture is the first AMD processor to support 512-bit AVX-512 vector instructions. We’ve already discussed their benefits (bigger or smaller) here. But the Zen 4 cores support another instruction set extension that used to be Intel’s pride and joy, and now the roles have reversed a bit: VNNI. It seems to bring huge performance improvements in a number of apps, despite the limited 256-bit width of Zen 4 SIMD units. Read more “Why is Zen 4 so fast in Topaz Labs AI apps? In fact it’s Intel’s doing” »

Raptor Lake is here: high speed with 7 Ultra node, no price hike

This week, AMD Ryzen 7000 processors and AM5 platform went on sale (we’ve already tested the 7900X). Intel will counter with their own new generation next month: the 13th generation Core “Raptor Lake”, which was unveiled on the same day, albeit with no physical availability yet. We now have definitive specs, as well as pricing. You’ll be pleased to know that most SKUs have not gotten more expensive, even though that was rumored. Read more “Raptor Lake is here: high speed with 7 Ultra node, no price hike” »

How good is AMD’s AVX-512? Does it improve Zen 4 performance?

When Intel Rocket Lake processors with AVX-512 instructions came out last year, we took the opportunity to test how they can improve performance in the x265 video encoder. Since the now-released Ryzen 7000 processors with Zen 4 architecture also support AVX-512, and there has been a lot of debate about the pros and cons with Rocket Lake, we went back to that test to see how AMD’s alternative implementation fares in it. Read more “How good is AMD’s AVX-512? Does it improve Zen 4 performance?” »

What RTX 4090 lacks: Nvidia RTX 6000 Ada is almost the full AD102

On Tuesday, Nvidia unveiled the first GeForce RTX 4000 GPUs with Ada Lovelace architecture. The fastest RTX 4090 SKU is based on the AD102 chip, but it’s already been confirmed that it’s a stripped-down version with 16,384 shaders, while the full chip has over 18,000 FP32 units. But Nvidia already has a card that has almost all of the units unlocked: the Nvidia RTX 6000 Ada, a GPU aimed at professional graphics and similar markets. Read more “What RTX 4090 lacks: Nvidia RTX 6000 Ada is almost the full AD102” »

Nvidia GeForce RTX 4000 cards are here: models, parameters, prices

Nvidia had their GTC 2022 keynote scheduled for Tuesday. As expected, company CEO Jen-Hsun Huang finally unveiled new gaming  graphics cards at this presentation alongside various enterprise, AI, and automotive stuff: the GeForce RTX 4090 and GeForce RTX 4080 with the new Ada Lovelace architecture. They’ll come to market in the next two months, however we don’t exactly have good news on the prices Nvidia will charge for them. Read more “Nvidia GeForce RTX 4000 cards are here: models, parameters, prices” »

End of an era: Intel to stop making Celeron and Pentium CPU

Yesterday we shared the news that EVGA is ending production of GeForce graphics with Nvidia GPUs after more than twenty years. Another “end of an era” moment has however happened virtually at the same time. Intel has announced that it will do away with the historic Pentium and Celeron processor brands that have survived since the 20th century. This is as part of image and branding changes the company is now planning for next year. Read more “End of an era: Intel to stop making Celeron and Pentium CPU” »

Zen 4 has a thermal boost, too. 5.85 GHz only at sub 50 °C temps

It is often commented on that Intel processors (like the upcoming Raptor Lake) have many different boost modes with various conditions, for example the temperature-dependent Thermal Velocity Boost. But it turns out that the Ryzen 7000s have something similar. According to unofficial information the top Ryzen 9 7950X model will also reach its maximum frequency only at low temperatures – impractically low, even. Read more “Zen 4 has a thermal boost, too. 5.85 GHz only at sub 50 °C temps” »

Cheaper Raptor Lake CPU will be just rebadged Alder Lake chips

Yesterday we reported that Raptor Lake processors are largely a continuation of the Alder Lake generation using the same ingredients, despite a large increase in achieved clock speeds and some architectural changes (cache). This is even more true than we thought. It turns out that only the higher-tier K-series models are actually new chips. The lower end, including all 65W Core i5s, are recycling silicon from the Alder Lake generation. Read more “Cheaper Raptor Lake CPU will be just rebadged Alder Lake chips” »

Intel Arc A770, A750, A580 desktop graphics officially unveiled

The coming of Intel Arc discrete graphics cards was strange – the company tried to keep up the audience excited with constant stream of news, but the efforts were thwarted by the long delays. But the long and slow, gradual trickle of all sorts of news bits is almost over. Finally, Intel has unveiled the gaming graphics cards with their most powerful GPU yet, up to 16 GB of memory and 4096 shaders. So now we’ll see what’s what. Read more “Intel Arc A770, A750, A580 desktop graphics officially unveiled” »

125W/65W/35W Intel Raptor Lake specs leak, cheap i5-13400 too

Lots of info has already leaked about 125W Raptor Lake processors that Intel is abotu to launch on October 20 as the 13th-generation Core for desktop. The leaked specs were not complete, however. Now, we have a document that clarifies virtually everything – the clock speeds of little cores, details of various boost clocks for different numbers of active cores. And even specs of some 65W and 35W models. Read more “125W/65W/35W Intel Raptor Lake specs leak, cheap i5-13400 too” »

Zen 4 architecture: chip parameters and IPC of AMD’s new core

AMD formally launched the Ryzen 7000 (with availability four weeks away though) and confirmed specs and pricing. The company has also disclosed a considerable amount of technical details and more has now leaked from unofficial sources. So let’s have a deeper look at Zen 4 architecture, the 5nm chiplets of the processors and finally the core’s IPC – which has actually improved more than the previously promised +8 to 10 % over Zen 3. Read more “Zen 4 architecture: chip parameters and IPC of AMD’s new core” »

AMD Ryzen 7000 announced: models, specs, pricing, availability

Today in the early morning hours, AMD held an event to officially unveil next generation desktop processors – the Ryzen 7000 with Zen 4 architecture made on 5nm manufacturing process. We learned the final specs of these processors and their new iGPU, as well as a store availability date and the prices the chips will cost. There’s been a surprise change in pricing in contrast with the previous leaks. In any case, Zen 4 is here. Read more “AMD Ryzen 7000 announced: models, specs, pricing, availability” »

Most Meteor Lake chiplets will be made by TSMC, not Intel

After the Raptor Lake processors due this year, Intel is preparing an ambitious new generation codenamed Meteor Lake for 2023. These chiplet-based processors and will be the first to use the 4nm process. However, it turns out that Intel is only making one of the four chiplets. The company that has always taken pride in its own factories will rely on its main competitor for most of the silicon in the Meteor Lake processor. Read more “Most Meteor Lake chiplets will be made by TSMC, not Intel” »

1 2 3 12