AMD Processors
Decrease font size
Increase font size
Topic Title: Trouble with Asus K8V, WD Raptor and OC'ing
Topic Summary:
Created On: 10/07/2003 02:16 AM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 10/07/2003 02:16 AM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

I have 2 WD600JB HDD's RAID 0 on the K8V's onboard Promise controller. A couple of days ago I bought a WD Raptor 36GB SATA drive to try. I noticed that it doesn't want to run on the Promise controller, only the VIA onchip SATA RAID. Also, the Raptor works fine as long as I don't over clock much. When I over clock to more than 215MHz FSB I get errors on the Raptor and it usually requires a reformat in order to fix the problem. The RAID 0 PATA drives have no problem at all hitting sync speeds up to 228MHz, or 240MHz async. Any thoughts as to what might be causing the problem? I am thinking it is related to the fact that there is no AGP/PCI lock.

Thanks for your help.
 10/07/2003 03:02 AM
User is offline View Users Profile Print this message

Author Icon
Dagalidis
Senior Member

Posts: 243
Joined: 10/07/2003

It seems that AGP/PCI lock will be a major problem when trying to OC the FSB Higher Than 215 - 220...
Lets hope a BIOS update can give us a future LOCK if mobo hardware support it...
Also multi selection is all what we need now to make a decent OC...

-------------------------
ATHLON64 3200+@2380(238X10) 1.644 Vcore Cooling By AMD Stock Heatsink ASUS K8V Deluxe (VIA K8T800 Chipset) 256x2 TWINMOS PC3700 5-3-2-2.5 at 2.8V Sound Blaster AUDIGY 2 Sound Blaster Megaworks Speakers 2.1 250D THX / 300 WRMS LEADTEK FX5900 128MB My ViVo
<span style='co
 10/07/2003 11:57 AM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

QUOTE (Dagalidis @ Oct 6 2003, 11:02 PM) It seems that AGP/PCI lock will be a major problem when trying to OC the FSB Higher Than 215 - 220...
Lets hope a BIOS update can give us a future LOCK if mobo hardware support it...
Also multi selection is all what we need now to make a decent OC...
Multi selection?
 10/07/2003 12:03 PM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

well, here's what comes to mind:

Via's implementation of SATA runs the data stream over the HT bus. When overclocking the HT "bus", you MAY be causing the Raptor to operate out of spec (i.e. SATA150 becoming SATA160, etc)... thus causing errors. does this make sense?

cheers,

dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/07/2003 12:28 PM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

QUOTE (Dave Graham @ Oct 7 2003, 08:03 AM) well, here's what comes to mind:

Via's implementation of SATA runs the data stream over the HT bus. When overclocking the HT "bus", you MAY be causing the Raptor to operate out of spec (i.e. SATA150 becoming SATA160, etc)... thus causing errors. does this make sense?

cheers,

dave
I guess. Do you think bumping the HT frequency down to 600MHz might help?
 10/07/2003 12:31 PM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

QUOTE (StormPC @ Oct 7 2003, 08:28 AM) QUOTE (Dave Graham @ Oct 7 2003, 08:03 AM) well, here's what comes to mind:

Via's implementation of SATA runs the data stream over the HT bus.  When overclocking the HT "bus", you MAY be causing the Raptor to operate out of spec (i.e. SATA150 becoming SATA160, etc)... thus causing errors. does this make sense?

cheers,

dave
I guess. Do you think bumping the HT frequency down to 600MHz might help?
this is a VIA board, right?

if so, returning the HT link to 600mhz (which is stock for nForce 150s; 800 is default for Via) would be a downclock...

post your stock specs and then your oc'ing specs here and we'll try to parse through them.

cheers,

dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/07/2003 12:54 PM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

QUOTE (Dave Graham @ Oct 7 2003, 08:31 AM) QUOTE (StormPC @ Oct 7 2003, 08:28 AM) QUOTE (Dave Graham @ Oct 7 2003, 08:03 AM) well, here's what comes to mind:

Via's implementation of SATA runs the data stream over the HT bus.  When overclocking the HT "bus", you MAY be causing the Raptor to operate out of spec (i.e. SATA150 becoming SATA160, etc)... thus causing errors. does this make sense?

cheers,

dave
I guess. Do you think bumping the HT frequency down to 600MHz might help?
this is a VIA board, right?

if so, returning the HT link to 600mhz (which is stock for nForce 150s; 800 is default for Via) would be a downclock...

post your stock specs and then your oc'ing specs here and we'll try to parse through them.

cheers,

dave
It's an ASUS K8V. Standard is 200MHz (400MHz DDR) and I like to run at 220-225MHz (440-450MHz DDR). I can do this no problem with a normal IDE drive or my Promise RAID 0 array, but the Raptor doesn't like it.

I normally leave the HT freq at 800MHz, but I was wondering if under clocking the HT would allow a higher FSB without causing problems for the Raptor?

I was also told that CRC flood should be enabled for better performance. Do you know if that is the case?
 10/07/2003 01:00 PM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

don't know anything about the CRC flood thing...I'll leave that to more experienced users to discuss.

Underclocking the HT (i'm assuming that there's a global clock for HT that affects memory AND peripheral support) should also decrease memory bandwidth (once again, IF there is one primary HT clock).

If not, I would assume that dropping to 600mhz would be ok to troubleshoot.

check it out and see if there's any drop in performance.

cheers,

dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/07/2003 01:07 PM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

QUOTE (Dave Graham @ Oct 7 2003, 09:00 AM) don't know anything about the CRC flood thing...I'll leave that to more experienced users to discuss.

Underclocking the HT (i'm assuming that there's a global clock for HT that affects memory AND peripheral support) should also decrease memory bandwidth (once again, IF there is one primary HT clock).

If not, I would assume that dropping to 600mhz would be ok to troubleshoot.

check it out and see if there's any drop in performance.

cheers,

dave
There is a loss of memory bandwidth for sure, and of overall performance but the A64 doesn't really need much bandwidth because of the efficiency of the onchip memory controller.

I might try playing with that global HT clock setting when I get home from work. If worse comes to worse I could pick up a PCI controller for it. I'm pretty sure that would take care of the problem. That way it's not running directly over the VIA's HT buss and I can just disable the VIA SATA bios completely. What do you think?
 10/07/2003 01:14 PM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

yeah, i'd see what the board has by way of HT speed manipulation...Effectively underclock a 16-bit HT link wouldn't do too much damage...

From what you described, it sounds like the HT clock is global, not independent for memory and peripherals...

goose it around a bit and see what happens...

cheers,

dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/07/2003 01:20 PM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

Yeah, I hate to buy a controller since the MBD already has 3 onboard HDD controllers, the VIA SATA, the Promise SATA/PATA and the normal IDE controller.

Thanks Dave.
 10/07/2003 02:39 PM
User is offline View Users Profile Print this message

Author Icon
Brian
Member

Posts: 98
Joined: 10/06/2003

i think it's probably the lack of a PCI lock that's causing this to happen


-------------------------
Folding For Team AnandTech Distributed Computing' ">http://forums.anandtech.com/categories.cfm?catid=39

My Humble Forums' ">http://forums.hellstormraiders.com
 10/07/2003 02:50 PM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

QUOTE (Brian @ Oct 7 2003, 10:39 AM) i think it's probably the lack of a PCI lock that's causing this to happen
could be, but since the VIA SATA is not on the PCI bus and the Promise is PCI (and the drive fails on both), it looks like a signaling issue or a hard drive issue, not a PCI bus limited issue.

cheers,

dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/08/2003 03:41 AM
User is offline View Users Profile Print this message

Author Icon
Dagalidis
Senior Member

Posts: 243
Joined: 10/07/2003

PCI/HTT LOCK is needed.....

-------------------------
ATHLON64 3200+@2380(238X10) 1.644 Vcore Cooling By AMD Stock Heatsink ASUS K8V Deluxe (VIA K8T800 Chipset) 256x2 TWINMOS PC3700 5-3-2-2.5 at 2.8V Sound Blaster AUDIGY 2 Sound Blaster Megaworks Speakers 2.1 250D THX / 300 WRMS LEADTEK FX5900 128MB My ViVo
<span style='co
 10/08/2003 12:51 PM
User is offline View Users Profile Print this message

Author Icon
StormPC
Senior Member

Posts: 548
Joined: 10/06/2003

Well I bought a PCI SATA controller last night and it didn't help a bit. Also, it's much slower than the K8V's onboard VIA SATA controller.

Looks like I'll have to use a regular WD800JB to bench Win2000 with.
 10/09/2003 05:41 AM
User is offline View Users Profile Print this message

Author Icon
Dagalidis
Senior Member

Posts: 243
Joined: 10/07/2003

QUOTE Also, it's much slower than the K8V's onboard VIA SATA controller.
This is BAD.....

-------------------------
ATHLON64 3200+@2380(238X10) 1.644 Vcore Cooling By AMD Stock Heatsink ASUS K8V Deluxe (VIA K8T800 Chipset) 256x2 TWINMOS PC3700 5-3-2-2.5 at 2.8V Sound Blaster AUDIGY 2 Sound Blaster Megaworks Speakers 2.1 250D THX / 300 WRMS LEADTEK FX5900 128MB My ViVo
<span style='co
 10/09/2003 06:33 AM
User is offline View Users Profile Print this message

Author Icon
Dave Graham
Senior Member

Posts: 351
Joined: 10/05/2003

QUOTE (Dagalidis @ Oct 9 2003, 01:41 AM) This is BAD.....
well it makes sense seing as how it's chained to the PCI bus...

i don't know what to tell you man....try running the HD diagnostics and see if the drive is actually bad...

cheers,
dave

-------------------------
----------
The opinions expressed above do not represent the views of Advanced Micro Devices or any of their affiliates.
-----------
Owner, Webmaster
Flickerdown Data Systems' ">http://www.flickerdown.com/stewks/
"The Fusion of Hardware and Life"
 10/10/2003 12:38 PM
User is offline View Users Profile Print this message

Author Icon
carrmic
Junior Member

Posts: 7
Joined: 10/10/2003

It's definately going to be a PCI clock issue. Changing the HT speed should not affect your memory bandwidth to the CPU at all. It will affect your bandwidth to the AGP slot and all the PCI devices. However 600 would still be plenty of bandwidth for most everything except maybe large AGP texture fetching.

The HT links are already running incredibly fast and I wouldn't expect much headroom especially on early boards. The expensive Gibabyte boards with 6-Phase power (GA-K8NNXP and GA-K8VNXP) may be better overclockers. However, many PCI devices are very sensitive to overclocking and don't tolerate it well. Keep in mind that SATA data transfers are incredibly fast right now too. 150 MB/s going over 6-pins instead of 80 for UDMA or 68 for SCSI is quite a lot. It's probably going to be very sensitive in early implementations too.

The best way to overclock would probably be with a multiplier at this point. Unfortunately that doesn't seem like an option at the moment.
Statistics
112018 users are registered to the AMD Processors forum.
There are currently 0 users logged in.

FuseTalk Hosting Executive Plan v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.



Contact AMD Terms and Conditions ©2007 Advanced Micro Devices, Inc. Privacy Trademark information