G    
 
Can the PA-2007 support the new 0.25 micron AMD K6 processors?
 
This site last updated
29th May 1999 01:30 AEST
(AEST= GMT + 10 hrs)
 
 Total number of visitors since 15 May 1998 
 
Home
[ Home ]
[ What's News ]
Hardware
[ Benchmarks ]
[ Motherboard
[ How-to's ]
[ FAQs ]
Other
[ Archive ]
[ E-mail me ]
[ Hot Links ]
[ My System ]
[ Your Page ]
  
  
This site is best viewed using 800 x 600 resolution
 
This site is designed to be viewed with Netscape 4.x,  
or Netscape 3.x.
 
 
 
 
If you have problems with this site.
 
E-mail me.
 
  
  
  
  
 
 PA-2007 

Technical Reference 
 alt.comp.periphs.mainboard.fic 
 
 
[What's News]
Saturday 29th May 1999 
  
  Good news the PA-2007 (rev 1.2) motherboards have successfully demonstrated they can and will support AMD newest processor, the 'sharptooth' K6-III.  At this stage FIC still do not endorse use of the K6-III processor on the PA-2007 motherboard.   
 
I have received plentiful e-mails about the cx4140 beta BIOS.  The cx4140 beta BIOS which adds support for the AMD K6 III still requires third party software utilities to tweak the cpu for maximum performance.  Here's a report from Mike and dreams do come true..... 
 
    Subject: K6-III on a PA-2007 
    From: mike@ibm.net 
    Date: 5/27/99 9:44 PM Eastern Daylight Time 
     
    FYI,  I have been running a K6-III 400 on a PA-2007 for 3 days now with no problems. 
     
    PA-2007 V1.2, 586A southbridge  
    Beta bios CD4140  
    412.4 Mhz (75*5.5)  
    2.4v   
    1 64MB SDRAM stick, manufacturer unknown. 
     
    My main worry prior to installation was the 12A+ current draw,  but so far,  so good.  The various components with heat sinks near the CPU only get warm enough to touch with a finger,  perhaps 110F. 
     
    Don't have the time to run benchmarks, but average Unreal FPS was up about 57% over a K6-1-225,  probably limited by my single Voodoo 2 card.  
     
    I haven't tried 450 Mhz yet.  I'm thinking that at some point I'll find the limit of the motherboard voltage regulator by melting something,  and I don't feel like installing and setting up a new MB right now.  
     
    And that's all I have to say about that. 
     
    Mike 
 
 
Will Friend sent me another report of a PA-2007 motherboard with a K6-III 412.5MHz effort: 
 
    Subject: PA2007 @ 412.5MHz 
    From: Will Friend 
    Date: Tue 25/05/99 9:19 

    Just thought I would let you know that I am running the PA-2007 at 412.5MHz with an AMD K6-III with no problems! 
      
    Configuration is as follows: 
      
    Beta bios CD4140b 
    256MByte ECC PC100 SDRAM 
    Front Side Bus @ 75MHz - CLK1 = 1-2, CLK2 = 2-3, CLK3 = 1-2 
    CPU Freq. Multiplier @ 5.5 - FREQ1 =1-2, FREQ2 =1-2, FREQ3 =2-3 
    CPU Voltage @ 2.4VDC - VR - 7-8 and 11-12 
      
    I am also using TweakBIOS and PowerTweak utilities to optimize various BIOS processor settings.  All in all,  I am getting VERY good performance!!  I think I'll keep this motherboard a bit longer! 

    I tried the 109CD12, 113CD13, and CD4140b BIOS.  All provide the same performance.  I am currently using CD4140b. 
      
    Thanks, 
      

    Will Friend

   
Good luck with your overclocking efforts and keep a close check on the hot components particularly the voltage regulator. 
 
 Friday 26th March 1999
 
  
FIC have surprised every PA-2007 motherboard owner with a new release BIOS update?  Thank you to all who sent e-mails to alert me to the fact.  There are two new BIOS's documented on the Taiwanese FIC website to add support for AMD K6-2 processor.  Has anyone tried either?  
  
The first BIOS ver 1.14cd14 dated 1999-03-17 is for the PA-2007 motherboard PCB rev 1.1 with VIA Apollo VP2 and South Bridge 586A only.  I think FIC messed up as I have download the cd1414.zip BIOS file (checksum: AD12H) and flashed my rev 1.2 Southbridge 586A motherboard and it reports the BIOS as version 113cd14 dated 01/04/1999-VT82C590VP2-2A5LC008C-00.  Windows 95B 2.5 didn't fuss and everything started as it should.  So far all benchmarks reveal no performance losses or gains - my system runs the same before and after.  May I also boldly state that the FIC documentation should read PA-2007 both PCB 1.1 and PCB 1.2 with VIA Apollo VP2 and South Bridge 586A only.  
  
The second BIOS ver 1.14cn14 date 1999-03-17 is for the PA-2007 motherboard  
PCB 1.2 with VIA Apollo VP2 and South Bridge 586B only.  I don't know about this one.  I need your help here again..... 
  
I hope to see benchmarks from anyone with a CXT core K6-2 or a K6-3 - does it add write allocation support?   
  
For anyone attempting a BIOS upgrade you should make sure have read the excellent information FIC have on the BIOS update procedure.  Check it out now.   
  
  
Can the PA-2007 motherboard really support a K6-3?  I have an interesting e-mail from Robert: 
 
    http://www.realworldtech.com/index.cfm    Posted a article  New Report: IIIrd Time may not be charm.  In this report it gave amperage requirements of the Pentium III and K6-III.  In the article it was stated that most voltage regulators for socket seven mobo's  were designed to handle ten amps,  and the K6-3 used up to 12.5 amps,  and to check your mobo before installing the new generation. I also tried to find this information on FIC/tw website to no avail.  Do you by any chance have this information, or is it something we should worry about?  Thanks again for a great website. Robert
The question is are the PA-2007 voltage regulators capable of supplying the needs of a K6-3 cpu running 300MHz and beyond without exceeding their ratings, overheating and frying themselves.  I don't know current ratings of the PA-2007 voltage regulators.  I need to check the markings on both of the regulators on the motherboard but that means stripping the board out of the case.  To be continued......  
 
 Sunday 7th March 1999
 
  
 Once again Frank T. Mars provides some interesting information about some software which could help us increase the performance of the AMD K6-3 cpus if we cannot get a new BIOS for the PA-2007 motherboard.  Below is the e-mail he sent me. 
 
    Hey Antonio, 
    The newer k6 chips..388..400 etc. use the new cxt core, which is good for about 8-9% in performance.  But you need a new bios update for this core And from what I can understand this Setk6v2 file can basically do the same....I think...and Hope 
    All this falls into place for the K6-3....same core...CXT. 
    For more info on the cxt core chips try www.anandtech.com new review of theK6-3.....Looks like it blows intel away...LOL 

    And for info on Setk6v2 try  www.lostcircuits.com  aka MS's BBS. 

    From what I understand this file setk6v2 is used when you bootup and you can preset the new chip features...new write features in the core. As soon as I can translate or find a english ver of this I'll send it to ya. But it looks like FIC will not give a new bios for the PA2007....So a file like setk6v2 or a file of this nature may be the only hope for running a k6-3 cxt core chip correctly on the PA2007 imho. 

    Regards 
    Frank T. Mars 
     

Thank you to everyone who has sent me e-mail in response to my recent update.  I have also had a note from Jim Williams who is currently successfully running setk6 version 2 with a AMD K6-2 333MHz processor running 375MHz on his PA-2007 motherboard.   He writes: 
  
    I'm with you about the K6-3. 

    I now have the K6-2 333 running at 375mhz on my PA-2007.   It has the CXT core and I use the SetK6 utility to enable write allocation.   It gets quite a boost from the SetK6 version 2 utility.   The original Setk6 doesn't work with the CXT core. 

    I'll also email FIC about the bios update.  If enough of us do it, maybe 
    they'll get off their butts and get one for the K6-3.   I think they owe 
    us that.  There are a lot of PA-2007 mainboards out there. 

    Thanks for your website! 

    Jim

  
Note:  I have not personally tried this software yet.   I hope to in the future as soon as the K6-3 are reasonably priced here in Australia. 

 


 Saturday 28th February 1999 
  
The AMD K6-3 has been released but will it work on the PA-2007 motherboard?  As so many have discovered the K6-2 processors work fine on revision 1.2 boards and can be overclocked most often without the aid of higher voltages.  AMD's K6-3 as well as the fastest K6-2 cpus have a new core known as the CXT core which include features to enhance its performance.  A new BIOS is needed to support the new K6 CXT core processors.  Some time ago I e-mailed FIC technical support but they say they dropped BIOS support for the PA-2007 a long time ago.  Here is the FIC Technical support response to my request for a BIOS update: 

"FIC decide not to upgrade PA-2007 to support K6(2) & K6(3) CPUs for a long time already.  It is our policy I can not submit your request." 
 

    Looks like unless we bug FIC Technical Support with e-mails we may never see a AMD K6-3 cpu working to its full potential on the FIC PA-2007 motherboard.
  

Monday 20th October 1998 
  
Leslie P. Chamness has e-mailed me some information about problems discovered with certain Western Digital hard drives and the VIA chipset.   

(Taken from the Western Digital web site) 

 "There is a known issue involving a small shipment of Western 
Digital high-performance WD Caviar Ultra DMA hard drives and the VIA VPX and 
VP2/97 (also marketed as AMD-640) chipsets. 

During the initiation phase of an Ultra DMA read data transfer, a series of 
commands are issued by the hard drive, the BIOS and the VIA chipset.  A 
deviation from standard protocol by the VIA VPX and VP2/97 chipsets is 
causing the system and the hard drive to go out of synchronization, 
resulting in a system that does not respond or locks up.   

Hard Drives Affected: 
 The specific Western Digital hard drives affected have CCC: C1 
firmware, were shipped between 9/10/97 and 10/8/97, with model numbers 
AC35100 and AC36400.   

Solution: 
 Current shipments of Western Digital hard drives are not affected by 
this VIA Chipset issue.  If using a Western Digital hard drive that has an 
issue with the VIA chipset, you must disable the Ultra DMA." 
 
 

     I had a WD hard drive with this problem.  I did not have a problem 
    with it until I upgraded my OS to Win98.  Then I thought it was a bug in the OS and worked on it with Microsoft  for about 6 weeks.  Then I came across the above message buried deep within Western Digital's web site.  The symptom of the problem was that my computer would crash at random whenever I enabled Ultra DMA.  Worked fine if DMA was disabled.  The solution was I contacted Western Digital and they exchanged my drive for another one under warranty.   Thought you might like to know... 

    Leslie P. Chamness 

  

Wednesday 30th September 1998 
  
A minor glitch has been discovered which affects K6-2 processors running Windows 95 OSR2, OSR2.1, or OSR2.5 (does not affect Windows 98) and clock speeds of 350MHz and higher.   It is not an actual silicon flaw but rather a software timing loop that is sensitive to the K6-2 processor speed.  The symptoms are "Windows Protection " errors after Starting Windows 95 and usually on reboot it does not occur again.  

There is more information on fixing the flaw available in a bulletin on AMD's web site.   

Thanks to the guys at HPC who have made the patch publicly available - Windows 95 patch - get it now before it is taken down. 

This patch should make it even easier to overclock the K6-2 300 to 366MHz when running Windows 95.  

     



Wednesday 9th September 1998 
  
Well it has been a long time between updates but unfortunately there has been very little earth shattering information about the PA-2007 motherboard to post.    
  
Thank you for your patience during my absence.  Don't ever forget this is your site too.  Now start up your mail program and send me whatever information you think is relevant so I can post it here.  
  
For all the US readers I have posted a link to an ftp site based in the US which is an exact mirror of FIC's Taiwan ftp site.  Thanks go to Jonny Gerow for sending me the url and giving US based readers a faster alternative to FIC Taiwan ftp site. 
FIC ftp mirror site 
 
FIC ftp mirror site - PA 2007 BIOS area 
 
  


Wednesday 5th August 1998 
  
Here is a tip which could help those of you trying to upgrade to Windows 98.  Thanks go to Nabil Yaghi for e-mailing the information to me.   
    I own an FIC PA-2007 motherboard running a K-6 233.  I had this weird problem when I installed win98 oem on my system.  The system would reset itself whenever i tried to load quicktime 3 or tried to access directx tool in the directx folder.  Also it would not access my network. The same system ran fine under win95b, c and NT4 with and without service release 3.  I contacted FIC technical support and they told me to enable USB in bios.  I did this and the system is running fine under Win98.  I just wanted to let you know as some of your readers might be having similar problems. 
  


Monday 27th July 1998 
  
Finally,   success with the EZ-S.M.A.R.T. software.  As reported earlier,  EZ-S.M.A.R.T.  is software from Micro House that will monitor and report the operating condition of S.M.A.R.T. hard disks.    

Today,  I installed a fresh copy of Windows 95b on the S.M.A.R.T. capable ide hard disk and then installed the EZ-S.M.A.R.T. software.  It does work,  and regardless of the BIOS S.M.A.R.T. HDD capability setting - whether enabled or disabled.  So, I must look to the FIC god to answer the big question.  What exactly does the S.M.A.R.T. BIOS support do?   

OK,  once installed the EZ-S.M.A.R.T. interface can be accessed by clicking on the system tray icon (near the time in the task bar).  The interface has five tabs, Drive Parameters,  Status,  History,  Configure and About.   Each time Windows 95 is started the EZ-S.M.A.R.T. software will monitor the S.M.A.R.T. hard disks for any S.M.A.R.T. triggers and alert you with an error message.   

Please note that it is important that the ide drivers are installed and working properly.  Only then will EZ-S.M.A.R.T. work.  I did not try the VIA ide bus master drivers as the EZ-S.M.A.R.T. FAQ recommends disabling the driver's ultra dma mode.  Ouch,  this will hurt performance.  
 

From Roger Thelemann I have some benchmarks. He has an  FIC PA-2007 with Bios 113CD135.  Running AMD K6-2 300 over clocked to 338, 75X4.5 at 2.4V 

Runs great and stable except for start up I have to hit the reset button for it to boot up properly. 

    My System 
    Windows 98 
    AMD K6-2 300 O/C to 338 
    FIC PA-2007 Rev. 1.2 
    Bios 113CD135 
    10ns 2x64 MB SDRAM (PNY) 
    Diamond Viper V330 PCI Video Card 
    Creative Labs 3D Voodoo 2 accelerator card 8MB 
    Western Digital HD  5.1 and 1.62 
    Sanyo CD 3 Changer 
    Creative Soundblaster AWE64 Gold 
    US Robotics 56K modem 

    Wintune 98 (offline test) 
    Clock rate   338 MHz 
    Dhrystone   876 mips 
    Whetstone   391 mflops 
    Overall mem   549 mb/s 
    Read mem   658 mb/s 
    Write mem   652 mb/s 
    Copy mem   283 mb/s  

    Winbench 98 
    Clock rate   338 MHz 
    CPUmark32   693 
    FPUmark   1090

 



Sunday 26th July 1998 
 
Now for a Quake 2 benchmark sent to me by the Funny Man:   

Just wanna drop yet another benchmark bomb on the already flooded with,  look at me and my 3DNOW scene.  One twist though.  I am NOT running a V2 and not running at 100mhz Super7 board. Sound boring.  I think not.  This was the most bang, banging?,  for my buck I ever got from a $100 computer upgrade. 
  
I tested at 68mhz (75 works but my CD-rom don't like it so i stayed with 68) and at 206/68 (get a baseline from my old k6-200),  274/68 and 308/68.  Well let the budget gamer masses know: This chip rocks when supported in UNREAL and Quake2.  
  
Quake2 3.17 
  

206/68 K6 206/68 3dnow 274/68 3dnow 308/68 3dnow 342/68 3dnow
rendition 512x200 21.9 24 28.5 31.6 33.4
rendition 640x480 21.3 23.6 25 28.6 30.3
powervr 640x480 13.1 13.8 14.3 15.8 17.2
vquake2 640x480 22.8 22.9 26.8
software 320x200 18.68 23.8 24.9 26.3 28.1
software 640x480 9.3 12 14.2 15.3 17.9
  
Default settings sound on,  
minigl 5/23 for stealth2 o/c to 59mhz,  
powervr default settings,  
software default,  
vquake AA=7, 
surface cache off. 

System.  
PA-2007 ver 1.2,  
32megs 10ns sdram,  
2.1gb seagate. 
  
NOTE: Without 3dnow was run with cl_paricles and l_dynamic off,  with 3dnow I turned it back on for full graphics and had very little performance hit as opposed to a lot without 3dnow. (loss .4fps with them off and 3dnow) 
  
This chip rocks!!  I have unreal scores too I will post them later today. 
  
FUNNY MAN 



  
Wednesday 22th July 1998 
 
I have delayed updating this site,  in the hope I would have the S.M.A.R.T. capability of the new BIOS 113cd13 working on my PA-2007 motherboard.  FIC tell me that you need software to access the S.M.A.R.T. hard disk diagnostic information.   

Sure,  when you look hard enough you find it.  I found EZ-S.M.A.R.T. v2.5 software from Micro House that will monitor and report the operating condition of S.M.A.R.T. hard disks.  The EZ-S.M.A.R.T. setup installs the utility and then launches,  smartupd.exe,  a Windows 95 SMART update program from Microsoft.  Now everything is S.M.A.R.T. capable but unfortunately I get the following error:  

 
 

Bad news.  I cannot solve the problem and I cannot prove that the S.M.A.R.T. capability works.  So is it the PA-2007 BIOS or the hard disk or the software (operating system included) that causes the failure?  I will have to contact Technical support - FIC,  Seagate and Micro House. 
 

A big Thank You to all those PA-2007 owners who have sent benchmarks.  I will post them soon,  so look out for them. 


Wednesday 15th July 1998 
 
Alexander Marx has sent me some interesting K6-2 3D benchmark results which I will post on the PA-2007 Technical Reference Benchmark Page.   Below is an excerpt from his note: 

    "I have been successful in getting the k6-2/266 up and running on my PA-2007.  

    My System: 
    FIC PA-2007 rev 1.2 
    Beta Bios 113cd135 cpu is identified as K6-2 
    128mb (2x64) SDRAM (Hyundai) 
    rem: I use fastest possible timings and 4-bank interleave 
    Standard Heatsink and Fan 
    cpu used is amd k6-2 266 2.2v 
    Clock generator on Board: W48C67-01H 1734T 
    Chipset used is VT82C590 Apollo VP2  (595_9737CD+586A_9727CD) 
     

    First i tried the standard 4x66 at 2.1v.  It ran perfectly,  but couldn't satisfy my appetite for speed in Unreal.  Then i went on to 4.5x66 at 2.1v.  Amazingly it ran rock solid.  This is my standard setting now. 
    I cannot approve that it works with any of the 75mhz settings because unfortunately my Sdram is not able to keep up (i think).  I also encountered this problem with my 166mmx p5, so it's not k6-2 specific.  The whole thing worked also with 5x66 at 2.1v but,  i had a few shell32.dll and kernel32.dll errors in win98,  so i tried 2.3v and it was stable until,  well... until my win98 blew up with a registry error that couldn't be fixed without a complete reinstall (it halted at startup).  I suspect,  although i am not completely sure,  that it was caused by overclocking.  I went back to 300/66 but changed the voltage to 2.2v as described on your page.  This Machine is running very fast now and best of all - rock solid. In five days of extensive (i mean it - 15 hours a day!) use i encountered not one crash or any strange behavior.  Lately i downloaded the optimized Q2 minigl from AMD and it rocks.  I want to give you some Benchmarks, too 

    If you are interested in more details,  I will be glad to give you more information to contribute to your page - It is the main reason that now I have pentium2-like performance and saved at least half my money.  By the way,  please excuse my terrible language - i am a german who knows english only from surfing the internet and playing silly computer games. 

    See you 

    Alexander"

  
[Content Highlights]
Hot news.   
ftp://ftp.fic.com.tw/bios/
  
Whoops my BIOS flash screwed up!  Go directly to the FIC site and check out the procedure on BIOS recovery. 
  
Hot news.  I have a new beta BIOS, 113cd135, which has solved K6-2 looping reboot problem for Frank T. Mars.  Thanks goes to Frank for testing this BIOS.  E-mail me if you have this problem and I can send you the BIOS file. 
 
Download the 113cd135 BIOS here!
 
Jonathan Sasse has put together an excellent report documenting his experience overclocking the K6 266 on the PA-2007 motherboard. 
  
I have done some more work and posted a preliminary How-to-guide for updating system BIOS.  It will be further enhanced when I gather more information on the common problems encountered when flashing the system BIOS.  E-mail me with your experiences and fixes. 

FIC have posted new memory support documentation for the PA-2007.  It documents the 64 MB and 128 MB tested with beta BIOS revision 113cd133.  The report is available on Your Page. 

The PA-2007 Archive page is online now.  You will find all the old news articles and more.  

Why does my VIA chipset-based motherboard crash under Windows98? 

Sorry to all owners of the PA-2012 motherboard.  I do not have a PA-2012 motherboard which I can do my core voltage testing with.  Therefore I cannot help you with a 2.2V setting. 
 


 
 [ Top ] 
 

 Copyright © 1998-99 Antonio Cataldo 
All Names and Trademarks are the rights of their respective owners. 
 Disclaimer 
 

  1