Tuesday, January 2nd 2018
VIA Making a Comeback to x86 CPU Market with Zhaoxin R&D Monies
The only other active x86 architecture licensee than AMD, VIA Technologies, is planning a comeback to the x86 processor market, bolstered by R&D investment by Shanghai Zhaoxin Semiconductor. VIA and Zhaoxin have been co-developing the ZX family of x86 processors for rollout in 2018, and at least on paper, the chips appear to have the chops to take on Intel's "Gemini Lake" SoCs. The new VIA-Zhaoxin combine CPU family begins with the KX-5000 "Wudaoku" SoCs launched late-2017. These are full-fledged SoCs, which completely integrate the chipset (including the southrbidge).
The KX-5000 chips feature 4 or 8 CPU cores without SMT, 2.00-2.20 GHz nominal CPU clock, 2.40 GHz boost clock, a dual-channel DDR4 IMC, a PCI-Express gen 3.0 root complex, an integrated graphics core, and platform I/O that includes SATA 6 Gbps, and USB 3.1 gen 2. This chip debuted on only one product from a major OEM, the Lenovo M6200 desktop model launched in China. 2018 could see a broader launch of VIA-Zhaoxin chips, with the KX-6000. While the older chips were built on the 28 nm process, the KX-6000 series will be built on the newer 16 nm process, feature 4 or 8 CPU cores clocked at speeds of up to 3.00 GHz, while retaining the feature-set of the KX-5000 series. These chips could realistically be touted as low-cost alternatives to Intel "Gemini Lake" SoCs, although Zhaoxin is making bold claims about its performance nearing that of AMD Ryzen processors.
Source:
Golem.de
The KX-5000 chips feature 4 or 8 CPU cores without SMT, 2.00-2.20 GHz nominal CPU clock, 2.40 GHz boost clock, a dual-channel DDR4 IMC, a PCI-Express gen 3.0 root complex, an integrated graphics core, and platform I/O that includes SATA 6 Gbps, and USB 3.1 gen 2. This chip debuted on only one product from a major OEM, the Lenovo M6200 desktop model launched in China. 2018 could see a broader launch of VIA-Zhaoxin chips, with the KX-6000. While the older chips were built on the 28 nm process, the KX-6000 series will be built on the newer 16 nm process, feature 4 or 8 CPU cores clocked at speeds of up to 3.00 GHz, while retaining the feature-set of the KX-5000 series. These chips could realistically be touted as low-cost alternatives to Intel "Gemini Lake" SoCs, although Zhaoxin is making bold claims about its performance nearing that of AMD Ryzen processors.
33 Comments on VIA Making a Comeback to x86 CPU Market with Zhaoxin R&D Monies
Holy @#$#
performance? value? less heat? cheap? embedded? small chip or something
if not i guess they gonna have hard time to enter the market and doing battle
The KT133 and KT266 ram issues, the 686B southbridge with UDMA mode issues and problems when tried to transfer more then 1GB from 1 drive to another.
via chipsets gave good performance, but at the cost of stability and issue after issue, after a while everyone who could afford to would escape to the AMD 760 chipset and those who couldnt would go for the Sis 756FX and 758 chipsters,
tell you the truth, Asrock wouldnt have become famous if it wasnt for their rock solid AMD sis 746FX and 748 chipset boards which replaced the Annoying and unreliable ECS K7S5a as the backbone of every Economy Athlon and Athlon XP build.
They did own S3 and Trident at some moment so i do believe they have some GPU knowhow, and they have all of Cyrixes all patents (including the canceled jashua which if released could have gone toe to toe with the P3 and K7). but its been too much time...
This issue you speak of only affected early pentium 4 chipsets, as the P4M266/P4X266 did indeed have memory related stability issues, but only when using more then 768MB of ram, and using that much memory was very uncommon in 2001.This issue (data corruption when transferring a file larger then 1GB from one partition to another - not more then 1gb of data from one drive to another) occurs only on early motherboards equipped with the VT82C686B. It's not really the controller's fault, but in fact a PCI latency issue. This issue has plagued select VIA chipsets from 00-01 and affected not only the HDD controller built into the SB, but addon cards as well, especially PCI sound cards and Gigabit LAN cards. There are BIOS updates and software patches available to fix this problem.VIA got a bad name in the Pentium II/III era when loads of cheap no-name slot 1 motherboards flooded the US (and europe to a lesser extent), and consumers unknowingly bought their fill. The problem there is not the chipset (VIA Apollo) but the poor quality and design of these cheap boards witch would be unstable to varying degrees (from unusable to tolerable). Slot 1 VIA boards from ABIT, Soltek and Asus will give any i440 board a run for it's money, and were cheaper. Vendors like PC-Chips, Matsonic, Jetway and so on released the aforementioned cheap and unstable slot 1 boards, and are the cause of the rumor. This can be also coupled with intel's anti competitive practices, where they would give large OEM's like Asus and Abit huge discounts to postpone the release of boards running competing chipsets, and release those in limited supply.
If you have the time and hardware you can see for yourself. Just go on ebay and buy the first PC-Chips/Matsonic/Jetway slot 1 VIA board you find, and try to build a machine around it. Then do the same with a proper VIA board like the Abit VA6 or Asus P3V4, and compare performance with an Intel 440 equipped motherboard. You will be pleasantly surprised.
There were some other issues as well. The aforementioned PCI latency bug, and poor chipset support in nvidia Forceware and Detonator video card drivers. There are some Detonator versions witch will cause BSODs on via chipsets under win98, but work fine on intel chipset boards.
Throughout my IT career and years of collecting hardware, I've only come across socket A SiS chipset boards inside cheap pre-buit machines with everything on board and usually no AGP slot. Performance and stability of these systems was abysmal. SiS did make good socket A chipsets but were too late to release them, and their reputation was already soiled by the plethora of aforementioned budget machines that flooded the low-end market. VIA was already offering the excellent KT400 by the time SiS pulled their shit together and released a stable / fast FSB200/DDR400 cabable socket A chipset - and guess what - they were still used in budget rigs - even tough this time around SiS had a winning product.
Same story with Super Socket 7 - SiS was late to the party, and most machines were using the MVP3 chipset from VIA.
AMD 760 boards are a rarity in eastern europe. I've only seen one, and that is the FIC AD11 in my collection. An excellent early DDR board, but from what I remember AMD chipset motherboards were nearly double the price of their VIA counterparts, so they sold really poorly. There was no motivation to buy one either, since you could get the same level of performance / features / stability from an equivalent VIA chipset board, and use the rest of the cash for more ram or a faster video card - witch is what most people did.
VIA had a bad name because it was underperforming compared with chipsets from Intel. I had a BX motherboard with my first Celeron CPU that I managed to burn out of stupidity(no experience with PC hardware). The replacement was a board with VIA chipset. Even with faster 133MHz SDRAM(BX could run the memory only at 100MHz) and AGP overclocking of the graphics card from 66MHz to 75MHz the VIA board was making my system slower than with the BX.