Intel 13th Gen Non-K CPUs Supposedly Up To 64 Percent Faster Than Previous Generation

Intel’s 13th Gen Raptor Lake CPUs and their performance shows how the chipmaker was able to retake the lead from AMD’s current Ryzen 7000 Series, in just one generation. Recently, a new leak seems to suggest that the brand’s non-K desktop CPU lineup is prepared to do the same, only this time, it’s against its predecessor.

The leak in question comes from a Twitter user who goes by the handle chi11eddog (@g01d3nm4ng0), and technically contains the performance metrics of the non-K 13th Gen CPUs. Specifically, they are graphs from Cinebench R23’s single-core and multi-core tests and more to the point, the tests were conducted on a total of five CPUs.

The non-K 13th Gen CPUs on the list are the 13900, 13700, 13600, 13500, and the 13400, which are essentially every CPU under the Core i9, i7, and i5 SKUs. On the multi-core test, the 13th Gen processors clearly have an edge over their 12th Gen Alder predecessors. The 13500, specifically, shows a performance improvement of up to 64% on the test, with the 2nd highest-performing CPU being the 13600.

On the single-core test, the 13th Gen’s 13900 was very clearly the CPU that has the most gains, with a 10% increase in performance over the last generation’s 12th Gen 12900. Overall, the worst-performing CPU on this list is the 12500, but that being said, it is still relatively decent. However, the 13900 of Intel’s 13th Gen Raptor Lake is expected to run 500MHz faster than the 12900 and also has eight more E-Cores than its predecessors. So, it really doesn’t come as a surprise that it would perform significantly better on the multi-core tests too.

As for when we can expect to see all five of these CPUs, Intel is expected to the launch its non-K 13th Gen Raptor Lake CPUs on 3 January, which is also when CES 2023 kicks off in Las Vegas. What makes them all that more appealing is that, while not overclockable, they all have a default TDP of 65W, nearly half of what their K-powered counterparts draw. At the same time, the chipmaker is also expected to announced its new B760 motherboard series.

AMD Supposedly Prepping Noise Suppression Technology

Rumour has it that AMD is supposedly working on an AI-based Noise Suppression feature that could be introduced in the near and immediate future. The alleged feature was first spotted on an official YouTube video on the brand’s channel, but has since been set to private.

Thanks to the quick hands of some fans online and through the magic and convenience of Reddit, a snippet of that video that “officially” shows off the alleged noise suppression technology. Again, it’s short and last for all but 40 seconds, but for the most part, it’s a sign that AMD is stepping up its game in the area of communication, as well as challenging a similar technology from its direct rival, NVIDIA.

We are, of course, talking about NVIDIA’s RTX Voice technology that the GPU maker launched close to two years ago, back when the pandemic was at its peak, virtually forcing a good chunk of the global population to work from home, until the vaccines became more widely available.

 Returning to AMD’s Noise Suppression technology and based on what we can see, the CPU and GPU maker is clearly working on something that works almost like RTX Voice, but with more tweaks and features made for Red Team, obviously. In other words, the feature should deliver a two-way noise reduction, cutting down on ambient sounds for both input and output.

Because AMD’s Noise Suppression allegedly uses AI-based algorithms to nullify these sounds, it could technically silence even construction work happening next door, thus ensuring a clear, crisp conversation between you and whoever else is on the other side.

As for its delivery vehicle: if AMD really is taking a page out of NVIDIA’s playbook, it could release its AI-based noise cancellation technology through the next major Adrenalin software drivers for its Radeon GPUs. The only other question is whether or not the brand will make Noise Suppression available only to RDNA2 cards, or if older Radeon GPU generations will benefit from the features. Although, all things considered, the latter seems unlikely, and that’s primarily due to these GPUs lacking certain internal components.

iPhone 13 Pro Supposedly Costs Slightly More To Build Than iPhone 12 Pro

As it turns out, an iPhone 13 Pro teardown has revealed that customers who bought the latest Apple release might be getting more value than those who bought last year’s iPhone. The higher estimated price is attributed to the more expensive A-series processor, NAND flash memory, display subsystem, and main enclosure.

The cost analysis by computer hardware specialist Techinsights estimates that the 13 Pro costs US$570 (~RM2383) for the base 128GB storage model, US$21.50 (~RM89) more than the US$548.50 (~RM2288) build cost for the 12 Pro.

Some of the components highlighted in the report are the A15 processor, the more compact TrueDepth array, and the display subsystem. The 13 Pro uses an LTPO panel used for the 120Hz ProMotion display, which might have added to the build cost when compared to the regular iPhone 13’s OLED panel.

The analysis also found that the A15 silicon found in the 13 and 13 Pro series might be the same processor. While Apple claims that the five GPU cores on the Pro are different than the four GPUs on the base 13, the tech giant might actually be disabling one of the GPU cores for the lower-end 13, as it did with the A12X chip. Techinsights says that both SoCs contained the same TMMU71 die mark and the exact same die size, which is 22.82% bigger than last year’s A14.

The price jump between the 12 Pro and 13 Pro is not as big of a leap as the one between the iPhone 11 and iPhone 12, which Counterpoint Research says may have costed 21% more due to the addition of 5G and OLED. The estimates are based on standard market rates so they might not be quite accurate given Apple’s supply chain scale.

The iPhone 13 Pro is launching at the same price as its predecessor, at RM4899. In contrast, the hardware researchers estimate that the Samsung S21+ 5G costs US$508 (~RM2124) to build and was launched in Malaysia for RM3999.

Intel Xe HPG Performance Supposedly On Par With NVIDIA GeForce RTX 3070

Intel has been keeping the lid tightly shut on all matters Xe related. The last time we heard anything about the chipmaker’s first discrete GPU was back in February, when Raja Koduri, its Chief Architect for Graphics, hinted at the Xe-HPG capabilities and its support for DirectX 12 Ultimate.

Fast forward to today and word on the wire is that we could be looking at a discrete GPU with performance on par with NVIDIA’s GeForce RTX 3070, if his recent tweet is anything to go by. Breaking it down, Koduri’s tweet shows two pictures from different time periods at Intel’s Folsom Lab. The takeaway here is the last bit of his tweet, where he proclaims that he is playing with a GPU that is approximately “20x faster” than a pre-production Crystal Well.

For those unfamiliar with that name, Crystal Well is what we know now as Intel’s Haswell CPU lineup; back then, the lineup’s integrated graphics was known as the Iris Pro 5200, which in turn was a part of the top-of-the-line CPU at the time, the Core i7-4770K.

While it may seem like Koduri is merely making a passing remark about Xe-HPG, there is actually some truth behind it. Thanks to PC World, we know that the Iris Pro 5200 had a UL Fire Strike’s graphics score of 1450 points. Multiply that by 20, and the number you’re looking at is 29000 points; that’s more or less on the level of the aforementioned RTX 3070.

That said, Koduri’s tweet is still, nevertheless, cryptic, and much of what we’ve talked about is still unconfirmed and mere speculation of what we can expect from Intel’s Xe HPG. So, until an official announcement by Intel, take this news with a large grain of salt.