The SIS 7001 USB controller on the P4S533 is utter garbage. It would crap out at any FSB over 140. This thing freakin blows. I went out and got a PCI USB 2.0 adaptor with 5 ports on it. Not only does it not have any compatibility or IRQ issues, its fast as hell (and USB 2.0), and runs fine up to 157 FSB. Before i couldn't go above 150 FSB without errors, and now i can almost get 160 FSB!! So i suggest you spend $20, disable onboard USB, and get yourself a nice PCI controller card.
Better stability by ditchin ONBOARD USB!!
- Thread starter OneOfTheseDays
- Start date
Sounds like my old kt266 chipset that wouldnt go above 138, but I could hit 150 easily. Swapped for a kt333 board and now I can hit 166mhz 
No point in ranting, really. The USB works as advertized at specified speeds, anything above is at your own risk, noone to complain to and noone deserving to be ranted at.
regards, Peter
regards, Peter
yea i know, but if everything was supposed to only work in spec, we wouldn't be able to overclock now would we? Anyways, my point is that the onboard USB controllers on many mobos simply suck. I suggest getting a new USB 2.0 controller with the NEC chipset, its nice and stable.
Anyone heard when USB controllers using the new NEC EHCI 1.0-compliant chip will be available? I'd be curious to know if the extra 20% of bandwidth makes much real-world difference with USB 2.0 peripherals.
So it sucks balls because it isn't able to run well at more than 13% overclocking?
Perhaps a better way of saying it is that the onboard USB actually works great, but you've found that adapter cards are more stable at higher bus speeds.
As for the speed, the USB controller is independent of the PCI bus (as is the IDE bus and anything else onboard these days), and on some boards, the PCI bus isn't directly clocked by the front side bus, but can actually be locked at normal speeds, or can have multipliers which allow you to overclock while still using normal PCI speeds. Don't know about the P4S533. Perhaps when you got it up to 166MHz the PCI bus has actually started using a different multiplier (x5) so that it is running at stock speed, whereas the onchip USB controller will be running at whatever the southbridge uses for a clock.
Either way, the controller is not garbage. It may not even be a stability problem for others, you may simply have a bad board for that particular feature. Other brands or even other Asus boards might work just fine with higher speeds and using the onchip USB.
Perhaps a better way of saying it is that the onboard USB actually works great, but you've found that adapter cards are more stable at higher bus speeds.
As for the speed, the USB controller is independent of the PCI bus (as is the IDE bus and anything else onboard these days), and on some boards, the PCI bus isn't directly clocked by the front side bus, but can actually be locked at normal speeds, or can have multipliers which allow you to overclock while still using normal PCI speeds. Don't know about the P4S533. Perhaps when you got it up to 166MHz the PCI bus has actually started using a different multiplier (x5) so that it is running at stock speed, whereas the onchip USB controller will be running at whatever the southbridge uses for a clock.
Either way, the controller is not garbage. It may not even be a stability problem for others, you may simply have a bad board for that particular feature. Other brands or even other Asus boards might work just fine with higher speeds and using the onchip USB.
actually the onboard USB controller is pretty bad. Even at stock speeds, some of my components would diasappear and then reappear. If you don't believe me go on the USBMAN.com forums and take a look.
Sudheer Anne is correct - USB hurts overclocking on most boards. just pop in a PCI USB card if you must have it.
functional onboard USB @ 175fsb? lol - yea right!
functional onboard USB @ 175fsb? lol - yea right!
well thugz i guess most people on this board don't seem interested in overclocking well. I mean who the hell says "The USB works as advertized at specified speeds" on a hardware tweakin forum? haha.... well if you have a mobo with usb 1.1 i suggest you spend the $20 and get a nice 5 port usb 2.0 card. The only thing holding me back now is the damn PCI/AGP divider, wish my mobo had it!!
I like the advice in this thread and will be looking for a USB 2.0 PCI controller this weekend because of it. Thanks.
"imts overclocking" is not equal to "sucks".
I have zero problems with the onchip USB on my K7S5A, which uses the SiS735 chipset with the same USB circuits as yours (technically not a USB "chip" since it's integrated into the chipset). I've never had any problems with it "losing" devices.
As for this being a hardware tweakers and overclockers board, yes, largely it is, but it also has a lot of people who are interested in the stock performance of parts, and who don't blame the manufacturer for their parts not working extremely out of specification. Most of us, if something doesn't work outside of spec, simply move on, or find another part that does work at the out-of-spec speeds or voltages. You however seem to think that failure to work out of spec indicates that a product is entirely unworthy of being used.
Unfortunately a lot of others think that too, which is why motherboards without extreme overclocking capabilities are nearly unheard of these days except from Intel and the very low end boards. Users now force a manufacturer to provide them with the means to run their systems out of spec, which then gives the users a chance to complain that their system won't work properly, which just serves to give the manufacturer a bad reputation, so they have to work hard to make their hardware work properly outside the specifications just so they don't lose sales, despite their product working EXACTLY as it should, within spec.
Once again, just because you can't crank a system way outside of specification and still get pure, clean, stable performance out of it, is not an indication that the product sucks. I've only used onchip USB with any system I've built for myself or others, and it always works. Whether it might limit my overclocking or not isn't a concern of mine YET, but when it does become a concern, I certainly won't blame the manufacturer for making a product that works as advertised.
Do you take the tires off your car to reduce the weight, then complain that the performance is sub-par?
I have zero problems with the onchip USB on my K7S5A, which uses the SiS735 chipset with the same USB circuits as yours (technically not a USB "chip" since it's integrated into the chipset). I've never had any problems with it "losing" devices.
As for this being a hardware tweakers and overclockers board, yes, largely it is, but it also has a lot of people who are interested in the stock performance of parts, and who don't blame the manufacturer for their parts not working extremely out of specification. Most of us, if something doesn't work outside of spec, simply move on, or find another part that does work at the out-of-spec speeds or voltages. You however seem to think that failure to work out of spec indicates that a product is entirely unworthy of being used.
Unfortunately a lot of others think that too, which is why motherboards without extreme overclocking capabilities are nearly unheard of these days except from Intel and the very low end boards. Users now force a manufacturer to provide them with the means to run their systems out of spec, which then gives the users a chance to complain that their system won't work properly, which just serves to give the manufacturer a bad reputation, so they have to work hard to make their hardware work properly outside the specifications just so they don't lose sales, despite their product working EXACTLY as it should, within spec.
Once again, just because you can't crank a system way outside of specification and still get pure, clean, stable performance out of it, is not an indication that the product sucks. I've only used onchip USB with any system I've built for myself or others, and it always works. Whether it might limit my overclocking or not isn't a concern of mine YET, but when it does become a concern, I certainly won't blame the manufacturer for making a product that works as advertised.
Do you take the tires off your car to reduce the weight, then complain that the performance is sub-par?
well i'm talking strictly about overclocking performance here. At stock speeds almost all onboard usb controllers are fine and do exactly what they are supposed to do and I think that's really great. But I'm only talkin bout overclockin performance here, and I don't blame the manufacturer because the product does work in spec. I'm only givin out some helpful advice to those who might be considering the same board and overclocking to higher fsb speeds.
well, it is utter garbage if it cant overclock
anything that cant overclock is garbage cause it doesnt have any headroom.
anything that cant overclock is garbage cause it doesnt have any headroom.
So when something works within specs...it's utter garbage.
Does your logic also apply to your Lexus, that can't do 0-60 in under 4 secs?
Didn't you put "good" ram in it so that wasn't a bottleneck? What's the difference?
John
Does your logic also apply to your Lexus, that can't do 0-60 in under 4 secs?
Didn't you put "good" ram in it so that wasn't a bottleneck? What's the difference?
John
when something works within specs it should have headroom. thats just plain good engineering.
your lexus can go faster then then the legal 55-75mph limit right? what if it couldnt? what if 75mph was all it could do?
your lexus can go faster then then the legal 55-75mph limit right? what if it couldnt? what if 75mph was all it could do?
Re-read the 2nd sentence of the start of this thread... "It would crap out at any FSB over 140"
So, it's got "headroom" above the stated specs. My 16yrs of engineering experience tells me that's 6+ over stated specs.
Thugs & Anne: Just be honest...You want something for nothing. I'm ok with that. But you aren't being objective in your analysis.
You didn't comment on the memory example. You acknowledge that putting "over-spec" parts in is ok (i.e. memory) but not others (i.e. USB). I would even be ok with your argument / whining if this limitation didn't have a cheap work-around.
In regards to the Lexus response: Having an RPM rev limiter vs. MPH limiter are 2 different things. I'm ok with an MPH limiter.
Just don't cry like little girls over $20.
So, it's got "headroom" above the stated specs. My 16yrs of engineering experience tells me that's 6+ over stated specs.
Thugs & Anne: Just be honest...You want something for nothing. I'm ok with that. But you aren't being objective in your analysis.
You didn't comment on the memory example. You acknowledge that putting "over-spec" parts in is ok (i.e. memory) but not others (i.e. USB). I would even be ok with your argument / whining if this limitation didn't have a cheap work-around.
In regards to the Lexus response: Having an RPM rev limiter vs. MPH limiter are 2 different things. I'm ok with an MPH limiter.
Just don't cry like little girls over $20.
RamIt
Senior member
- Nov 12, 2001
- 777
- 186
- 116
Garbage to me might be gold to another. If onboard USB is unstable when overclocked I consider it not worthy of purchase.
Quote from ASUS:
Extreme Overclocking
ASUS P4S333 offers robust overclocking options to maximize your system performance.
Easy DIP Switch or Jumperfree? overclocking from BIOS.
Flexible CPU Core Voltage Adjustments in 0.05V increments over defaults
Accurate 1MHz adjustments of system-bus-frequency
Overclockable AGP bus
Adjustable VIO memory voltage
Adjustable FSB/PCI/MEM ratio
Rock-solid stability
Quote from ASUS:
Extreme Overclocking
ASUS P4S333 offers robust overclocking options to maximize your system performance.
Easy DIP Switch or Jumperfree? overclocking from BIOS.
Flexible CPU Core Voltage Adjustments in 0.05V increments over defaults
Accurate 1MHz adjustments of system-bus-frequency
Overclockable AGP bus
Adjustable VIO memory voltage
Adjustable FSB/PCI/MEM ratio
Rock-solid stability
And for $20 the problem goes away.
It's important to note I never said the information that started this thread wasn't good. I actually find it valuable. I could just do without the whining.
OCers are a funny breed. Fact: Every board has limitations. Whether those limitations are a material matter to you, is the question. If this USB issue rules out the board for your application, ok. But that does not constitute "utter garbage".
Personally, I hate USB, it's too temperamental -- brand-name devices not playing well together, etc.
It's important to note I never said the information that started this thread wasn't good. I actually find it valuable. I could just do without the whining.
OCers are a funny breed. Fact: Every board has limitations. Whether those limitations are a material matter to you, is the question. If this USB issue rules out the board for your application, ok. But that does not constitute "utter garbage".
Personally, I hate USB, it's too temperamental -- brand-name devices not playing well together, etc.
Sudheer,
BTW, when your overclocked USB crapped out, was the onboard USB controller configured to interface on the embedded MuTIOL bus or the PCI bus?
BTW, when your overclocked USB crapped out, was the onboard USB controller configured to interface on the embedded MuTIOL bus or the PCI bus?
What do you mean configured to ride either bus? The onchip controller is embedded in the processing of the southbridge, rather than being a separate process that interfaces through the PCI bus. This isn't something you can configure.
As for it being "Extreme Overclocking", they never claimed that all functions will be available if you overclock, and in fact all motherboard makers specifically state that if you overclock, they can't support the system. To quote from Asus's manual for this board:
You're whining about a non-issue. You choose to try to get something for free, namely performance beyond the rated specs, you can't complain if it doesn't work entirely stable. You CAN state your experience with it, but it's still just your opinion that the product is "garbage" because it works not just at stock speeds, but at significantly overclocked speeds, but not quite overclocked enough for you.
As for it being "Extreme Overclocking", they never claimed that all functions will be available if you overclock, and in fact all motherboard makers specifically state that if you overclock, they can't support the system. To quote from Asus's manual for this board:
You're whining about a non-issue. You choose to try to get something for free, namely performance beyond the rated specs, you can't complain if it doesn't work entirely stable. You CAN state your experience with it, but it's still just your opinion that the product is "garbage" because it works not just at stock speeds, but at significantly overclocked speeds, but not quite overclocked enough for you.
What do you mean configured to ride either bus? The onchip controller is embedded in the processing of the southbridge, rather than being a separate process that interfaces through the PCI bus. This isn't something you can configure.
I don't know how many 645 BIOSes you've seen, but every 645 Award BIOS I've seen (including the Soyo P4S-645DX, Abit SD7-533, Gigabyte 8SR533, and Epox 4SDA off the top of my head - but no Asuses) has selections called "USB0 Interface Acesss" and "USB1 Interface Access" (there's also an "IDE Interface Access" option as well). The two options for each selection are "PCI Bus" and "Embedded Bus." I've never tried changing those selections myself; that why I asked Sudheer.
BTW, you should notice there was a pattern there - every PCI-like device on the MuTIOL bus has those selections available if the board implements them. For example, the Soyo P4S-645DX which uses the 961's SiS 9xx MAC has a "MAC Interface Access" selection. The Soyo apparently also allows use of the 961's AC97 audio instead of the C-Media 8738 so it has an "Audio Interface Access" selection. Each selection has the PCI/Embedded Bus options.
I don't know how many 645 BIOSes you've seen, but every 645 Award BIOS I've seen (including the Soyo P4S-645DX, Abit SD7-533, Gigabyte 8SR533, and Epox 4SDA off the top of my head - but no Asuses) has selections called "USB0 Interface Acesss" and "USB1 Interface Access" (there's also an "IDE Interface Access" option as well). The two options for each selection are "PCI Bus" and "Embedded Bus." I've never tried changing those selections myself; that why I asked Sudheer.
BTW, you should notice there was a pattern there - every PCI-like device on the MuTIOL bus has those selections available if the board implements them. For example, the Soyo P4S-645DX which uses the 961's SiS 9xx MAC has a "MAC Interface Access" selection. The Soyo apparently also allows use of the 961's AC97 audio instead of the C-Media 8738 so it has an "Audio Interface Access" selection. Each selection has the PCI/Embedded Bus options.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter