|
|
Advertisement:
|
|
OCAU News |
I'll start tonight with a few more rants.. If you don't care about the possible performance benefits of PCI/AGP bus overclocking, or about KT133E then skip these two big paragraphs.
First is about the AGP/PCI overclocking performance increases I asked for feedback on a few days ago. I've got some really decent emails from a few of you, so thanks to everyone that emailed! I've basically come to a few conclusions. Increasing the BUS speed is not at all similar to overclocking the core/ram settings of a video card. The closest performance difference between a standard and an overclocked AGP bus is best compared to the difference between AGP 2X and AGP 4X settings. The different AGP X ratings are im theory similar to what we know as DDR ram and the Quad Pumped FSB the P4 uses. AGP 1X is normal one-transfer-per-clock speed, 2X AGP is similar to DDR transfer speeds and 4X agp is double-DDR, or QDR. So now that you know what it means, how is it related to an overclocked AGP bus? Well, 4X AGP is "effectivley" 4x66. So if you up the AGP bus, it'll have the same effect as going from 2X AGP to 4X AGP, except on a smaller scale. Most of us should already know that going from 2X AGP to 4X AGP is a negligable performance increase - even though it ~theoretically~ doubles the available bandwith for the AGP card to exchange data with the CPU/RAM subsystems. So an overclocked AGP bus won't bring much of a performance benefit because there isn't bottleneck. If textures become massive, then maybe it will. But for now the only benefit an overclocked AGP bus possibly brings is the tandem affect of an overclocked RAM bus. That solves that.
Next, KT133E. I havn't seen any tech info on what this chipset is, and it's had me confused for a while. I think I've got it figured out now. Basically, it's the exact same chipset used in mobile Athlons/Durons. I take this assumption from the PowerNow! support. The good overclockability of the chipset I simply attribute to Via getting better at making KT133 chips. Newer KT133A's overclock better than old ones, as do newer thunderbird chips (AXIA? pfft.. AZIA baby!), so thats possibly why it has no qualms running at fairly out of spec speeds. But why are they selling KT133E's when KT133A does effectively the same thing? Two theories.. the first is that it's uneconomical just to make the small number of chipsets demanded by the laptop market, so they make a large amount and palm the leftover ones off to the desktop market. Nothing wrong with that. The other theory is that the chipsets going into laptops are sort of like GF2U chip's.. ie: binsorted high quality ones. Except they're not looking for overclockability, they're looking for low power consumption and low heat output. That would explain the low cost. So what does this mean? KT133E is effectivley the hot running or power hungry chips that wouldn't cut it in a laptop. No problem with that. It also ~possibly~ means that with the right BIOS you could ~forseeably~ FSB overclock a laptop. Heat limitations not withstanding. That'd be pretty cool.
I promise the rest of the news won't be so long winded. If rants like that totally bore the bejesus out of you, let me know. Likewise if you're interested in what I've got to say, let me know. here. Oh, and thanks to everyone who gave me motivation for uni. It helped alot. Dosn't mean the assignments finished yet tho :/
Manaz, man of the long email disclaimer, pointed out this fairly comprehensive timeline showing the development of Unix and other *n?x type OS's. If nothing else, you can at least read it and know where Linux came from. Also on Unix news from Manaz.. check out this page, hosted on a Microsoft Server no less. Ummah.. look who is NOT respecting their competition.
The US government is throwing heaps of money at students studying computer science and other IT-type degrees on the agreement that they come work for them once they're done. The have finally figured out that they don't have as many IT security heads as they should.. I wonder until Mr.Alston figures that out. Thanks to Dan Cooper for the link.
Rizenet has been busy behind the scenes, but has finally come back to the foreground and updated their CaseMod gallery. I dun reckon it's up to PCDB standards tho.. what do you reckon? NOT you MWP :D
2CPU have some info up about support for Via chipsets under Win2K with SP2. To cut their short story even shorter, they conclude that Win2K SP1+4in1's = SP2. In other words, SP2 finally brings decent Via support to the Win2K codebase. For SMP Pro133A boards at least. No word on KT133 boards or 686B though.
Icrontic has taken a squiz at Version2 of CPUfx's Core cooler. It's a pretty pricey thing, especially for an aliminium air cooler. At least it comes with a tube of Artic SilverII.. even tho, is it worth the price?
Another hs/f review, this time the OCZ MonsterII on TacoNuts.
Final hs/f review on PCExtremes.. an ADDA B53. Whatever it is, I dont know. It's big tho.
Tech Report have a new writer, and the new writer has a new article about hardware T&L. A pretty decent point is made - has T&L been worth it? nVidia made lots of noise about T&L being the Next Big Thing, but I havent seen much of it. I dont really feel we're all that much better off because of it. In the day of GHz+ CPU clocks is hardware T&L anything more than a marketing ploy?
The rounded cable craze is spreading.. ClubOC have reviewed a pre-rounded 68-pin SCSI cable.
Return to OCAU's News Page
|
|
Advertisement:
All original content copyright James Rolfe.
All rights reserved. No reproduction allowed without written permission.
|
Advertisement:
|
|
|
