KT266 boards are almost here !

[pvsurfer]

Like a lot of you, I've been looking forward to the new motherboards based on VIA's KT266 chipset. So I thought you might be interested in what I've found out so far...

I'm pretty certain the first DDR Socket-A motherboard based on VIA's KT266 chipset will be the A7V266 from ASUS. It will have two 184-pin DIMM slots for PC1600/PC2100 DDR SDRAM, three 168-pin DIMM slots for PC133 SDR SDRAM, 5 PCI slots, 1 AGP Pro slot and 1 ACR (yuck!) slot. It will also come with an integrated four-channel sound controller from C-media and an ATA/100 RAID controller from Promise. At this time, availability is projected to be the right around the beginning of April. MSI is saying that it will ship its KT266 board in April as well.

By far, the most impressive feature of VIA's new KT266 chipset is its newly designed North (host) and South (client) bridges, the VT8633 and VT8233 respectively and the V-link bus that connects them at a true DDR of 266MB/sec. Since the North and South bridges of all of the current Socket-A board (including the DDR boards) are connected by a PCI bus operating at 133MB/sec., you can see that the new KT266 chipset doubles the prevailing host bridge to client bridge communication rate and that should (theoretically-speaking) make a noticeable improvement in overall performance!

That's all I know right now, but if anyone has something else to add, your input is welcomed.

 

[Guest]

Do you know anything about the MP boards that are about to come out? I understand thw wont be hypertransport but can we expect similar to the KT266?
Waiting...
Anim88tor

 

[sinner]

-shakeshead
I really jumped the gun with this KT7-R @ christmas time, didn't I, lol. Shoulda waited a few months. I got win98se and win2k dualbooting, for some reason 3dmark2000 always locking up in win98se now. odd
didn't before I loaded win2k. odd

=]

-sinner.

 

[Guest]

I am planning on purchasing a 1GHz Athlon-C CPU with the A7V133, and 256MB PC133 RAM. Now here is my question. Would it be a better idea to wait a month and get a A7V266 with 128MB of PC2100 DDR? How much more would it cost to get a A7V266 with DDR? Thanks

 

[pvsurfer]

I don't think anyone can tell you if you would be better off waiting, but my feelings are that if you are building/buying a system including new memory, why not invest in the latest technology?

Predicting prices a month from now is anybody's guess. But of course, the DDR system will cost you somewhat more (perhaps on the order of 25% more for like parts).

 

[pvsurfer]

Sorry Anim88tor, but I can't help you (I'm not sure I know what an MP board is).

 

[Guest]

Sorry, The 760MP, dual processor chip that is reportedly about to be released.

 

[Vorg]

>>and 1 ACR (yuck!) slot. It will also come with an integrated four-channel sound

I also hate that ACR thing. I also don't like built in sound. The only built in stuff shoule really be the com and parrallel ports. Leave the rest plug in and PROVIDE THE BLASTED SLOTS!! so WE can decide which sound system and if we want raid or what ever.

At least stop putting the sound and ARM crap on it and go back to 7 slots. PC's started with 5 slots, went to 8 and now are headed back down to 5. They just got a new computer where my friend works and it only had 3 slots. What's more, they needed to put 2 ISA cards in for data monitoring and it didn't have any ISA.

 

[pvsurfer]

Since I much prefer listening to music on my stereo system (while sipping some wine in the comfort of my living room), I actually like the idea of an integrated sound chip on the mobo, which is good enough for my PC listening purposes. On the other hand, I don't understand why a lot of mobo mfrs go with AMR and CNR slots (don't know anyone who uses either), or integrated video (I haven't met anyone yet who would settle for integrated video over a good video card). That added "junk" results in higher mobo costs and end-user prices!

 

[pvsurfer]

Of course (I should have known that anim88tor)... This is from AMD's website:

Question: What is the development status of the AMD-760MP chipset? When does AMD plan to introduce a multiprocessor platform?

Answer: We are publicly demonstrating the AMD-760MP multiprocessing chipset, and expect multiprocessing AMD Athlon systems will be available during the second quarter of 2001.

Frankly, I just don't see the cost-effectiveness of multiprocessing ON THE DESKTOP. There is very little multithreaded software to take advantage of MP and while Win2K supports MP and will make (limited) use of it, I just don't see how you can justify the extra cost of the mobo and 2nd processor.

 

[Guest]

I do highend 3D so I can sue as many fast processors as I can get. For anything non multi threaded The 760MP doesent have any advantage.
Anim88tor

 

[Guest]

If you want the best system, than AMD and some other listed components would only be considered in the middle to low end category.
For a reasonable amount , you would be able to get the following-------
ASUS PENTIUM 4 1.5 GHZ MB w 1.5 GHZ CPU

2x 128 Meg 800 mhz RAMBUS (this is dual channel for 3.2 GBPS bandwidth 2 times DDR !)

2x 15,000 rpm Cheatah 18 GIG Ultra 160 scsi Drives (3.8 MS
one for OS and one for software, can operate concurrently for close to 180 MBPS sustained ! )

Geforce 3 64 meg video

Kenwoood 72X Multilaser CD or Toshiba 16\48 DVD
19" Viewsonic P95f
CASE -COOLER MASTER 201 www.coolermaster.com ALL RAW ALUMINUM GIANT HEATSINK !

WINDOWS M.E with DD3d 8a both P4 compiled !!

this will outperform any AMD system with similiar components
RAMBUS has superior maximum bandwidth expecially dual channel where both ram modules are effectively operating at 1600 mhz concurrently

also with P4 compiled software like dd3d 8a, there is a trememdous speed gain.
P4 has dual FP units operating a 3 GHZ or 2x, and can handle more operations
also P4 has a cache prediction rate of 95% thanks to a table that is 8x times larger than P3 or AMD
it has a 20 stage pipeline and 8 way cache as
well as a 400 mhz memory bus compared with 200-266 for AMD

any 3D app using P4 optimization will outshine anything
the spec mark for the P4 is 550 and its FP is 540 !
currently the fastest CPU around
do yourself a favor get the P4, it will be upgradable past 2 GHZ this year !
we build thousands of high end systems for clients and test alot of hardware like TOM does, so I am speaking from hands on knowledge

The P4 is the fastest we have tested by far

CYBERIMAGE
<A HREF="http:// http://www.4CyberImage.com " target="_new">http:// http://www.4CyberImage.com </A>
Ultra High Performance Computers-

 

[IntelConvert]

CYBERIMAGE: I sure don't want to start another needless Intel-AMD bashing contest here, but I'm afraid you are going to hear from a few others...

Frankly, I've got to tell you that while perhaps in theory you are correct, in the real world, I don't think so.

Over the past several months, I've read just the opposite here on Tom's Hardware as well as on AnandTech, Ace's Hardware and just about every other popular website. So I got to ask: Where are the independently-run comparative test results that back up your statements?

 

[Guest]

CYBERIMAGE: Regarding P4 and Geforce3, I don't know this to be true but the following link reports that the Geforce3 isn't working with the P4:

Scary quote:
"According to eyewitnesses, a Pentium4 system with GeForce3 is quite unstable, in some cases, even dynamic test could not be passed"

The link:
http://www.reactorcritical.com/index.shtml#l658

 

[Guest]

ahh that would be because TOM rarely listens to manufacturers reccomendations especially INTEL's when testing their equipment becasue he is biased and hates them for the past fiasco about them suing him.

you have to use optimized code for P4, or then use unoptimized code for AMD. he and others use
non sse2 optimized code for the P4 agains 3d NOW optmized
AMD code..
total BS..

many tests are available independantly from pro organizations, that have more credibility, like
SPEC ORG
the spec mark is the workd standard for comparing CPU accross all platforms, Alpha, SUN, AMD, SGI, INTEL
and it shows systems based on the P4 to be the highest score yet some 550 compared to AMD 426 !!!
that is HUGE

PC magazine shows that when using 3d optimized code for P4
the P4 is loads faster in games, MP3, Mpeg coding,rendering etc..
they even admit that they first tested with dd3d 7 and when the reveived dd3d 8 they changed the article because it game a hugh difference..

you have to understand cpu architecture to know the P4 is superior..

the bandwidth and performance of P4 and RAMBUS is higher no one disputes that. that is a fact.
what they say is that there is no visible difference..
well I ask you,
if you open NOTEPAD, for example on a 400 mhz CPU then on a
1000 mhz CPU you will not see the difference either
WHY
because you did not saturate the CPU with enough load to show the differences

Mpeg, Mp3 encoding, rendering, 3d games hammer the CPU
showing the P4 advantages..

a bigger glass does not hold more water on appearance
than a smaller one if you put the same amount of water in them.. FILL THEM and you then see the diffference..

Multitask, run background tasks, rendering, encoding etc..

NOT running Office apps or siple looped benchmarks that
are not optimized in any event..

links

http://www.rambus.com/products/products_benchmark_main.html#1

http://intel.com/procs/perf/Pentium4/index.htm

http://www.zdnet.com/special/stories/sc/desktops/reviews/0,12492,2655797,00.html

there are too many to list here, but you just remember you have to saturate the CPU's to get a accurate result and you have to allow P4 to have optimized software otherwise
it cannot use its risc-cisc translators and decode them into microops properly and it stalls..
WINDOWS XP and Office XP will be released in 3 months
I am running them NOW and they are alot faster still

hope that clarify's things..



CYBERIMAGE
<A HREF="http:// http://www.4CyberImage.com " target="_new">http:// http://www.4CyberImage.com </A>
Ultra High Performance Computers-

 

[phsstpok]

Have you tried the <A HREF="http://www.amd.com/products/cpg/athlon-duron/amd_win2k_patch.html" target="_new">patch</A> for the AMD/Win2K AGP problem?

 

[IntelConvert]

Kind of hard for me to accept that most of the major PC hardware sites are intentionally biasing their articles towards AMD's favor.

Granted RDRAM has excellent bandwidth, but (reportedly) very high latency!

Finally, RAMBUS and Intel links hardly qualify as being independant!

I have an open mind about this, but I would have to see independant "real-world" comparative tests (that are NOT optimized to take advantage of SSE2, as that's not playing fair).

 

[Guest]

actually the INTEL and RAMBUS tests were done independantly by 3rd parties..

they thought of that alreeady )

latency is not really relevant when you put data on both sides of the rising and falling side of the signal,
as rambus does, and besides, throughput and bandwidth is what limits most pentium 3 and AMD systems, not latency...

also RAMBUS just showed their new RAMBUS 2 which will be 1600 mhz and give up to 6.4 GPS bandwidth...

also having 2 concurrent ram channels operating in sync is
obvious to any layman to be superior..

again you just have to saturate the computer to show the difference in raw data handling..

CAMERON

CYBERIMAGE
<A HREF="http:// http://www.4CyberImage.com " target="_new">http:// http://www.4CyberImage.com </A>
Ultra High Performance Computers-

 

[Guest]

Yeah well maybe I can help you with some inside info you did not know.

ANANDTECH is run by a 16 year old from his bedroom who started the webpage because he had to get a replacement cpu for his piece of [-peep-] cyrix CPU and is a hobbyist at best.

and TOM favors AMD becasue INTEL threatned to sue him for publishing false and misleading data about a alpha engineering sample that was not properly reviewed..

If I were you I would not site them as unbiased sites!

CAMERON

CYBERIMAGE
<A HREF="http:// http://www.4CyberImage.com " target="_new">http:// http://www.4CyberImage.com </A>
Ultra High Performance Computers-

 

[yoda271828]

LOL, You ought to get along with FUGGER and AmdMeltdown really well.