Abit IP35-E Review (500MHz FSB board)...$90 @ NewEgg + $6.61 ship

[hokiealumnus]

Originally posted by: RetailDirect
Sorry to keep re-posting but I just realized that CPU-Z has my CPU voltage at 1.392~1.400 when I have it set to 1.45 in BIOS.

Whats up with this?

Also, I never did see a place to change from 2T to 1T in BIOS for RAM timings.

Thanks for the help!!

This board under volts a little bit and if you're under load, that's increased by Vdroop. In my experience, disabling C1E & EIST reduces the undervolt a bit. If overclocking, doing that will get your actual voltage closer to what you've set.

The 2T/1T setting was added with BIOS v.12 (well, actually with beta13, which was released before 12, but they kept it in 12).

 

[monohouse]

I have to say thanks to SerpentRoyal for his accurate information regarding this board, I bought it almost only because of it, and the results ?

I was able to achieve 400 mhz fsb (from 333) without changing anything else in the bios ! no voltage of any components ! maybe it's because of my cooling but doesn't matter I think SerpentRoyal isn't right just about the best price/performance ratio of this mobo, but maybe also it is THE best mobo for the hardcore ! I have no regrets from the absence of RAID or 1394 or any more SATA channels never used any of those anyway (need UPS to prevent major data losses and have no 1394 hardware regardless of the fact that 1394 hardware is expensive), and I only have one SATA channel in use anyway.

yes the NB is hot, BUT if you put good grease on it and on top of that grease you put a water block .... let's just say that so far I have not seen ANY NB temperature changes (during full load or not), it simply stays at room temperature and that's it, the only thing that does change in temperature is the CPU, which rises about 15C during full (dual) load above room temperature (measured with latest beta version of coretemp, and the NB temp was measured with latest beta version of everest).

so from 2660 mhz to 3200 mhz without even starting to mess with any voltages (but I will get to that later), it's a 540 mhz increase, there were times that (with my old Mobile 2500+ Athlon XP) I needed to fight to get that mutch, eventually I did (from 1800 to 2333 mhz) but the FSB of the mobo (KT333) was what stucked any further increases. and with this board more than 500 mhz without any efforts is just amazing.

my ONLY complaints about this board is: high memory latency (200 cycles) and double post during boot

 

[SerpentRoyal]

Originally posted by: RetailDirect
Sorry to keep re-posting but I just realized that CPU-Z has my CPU voltage at 1.392~1.400 when I have it set to 1.45 in BIOS.

Whats up with this?

Also, I never did see a place to change from 2T to 1T in BIOS for RAM timings.

Thanks for the help!!

If you've already added VTT, NB, and SB, then you need to run Memtest86 test #50 to test RAMs, and raise Vcore. It's normal for Vcore to drop 0.05V from BIOS when you stress the CPU.

 

[mike525]

SerpentRoyal,
why do i need to bump VTT, NB, and SB up one notch with FSB above 450MHz? I currently have my vCore at 1.4525 and my fsb at 475. Memory timings are relaxed and core temps are staying below 60C under load. In the bios, there was only one other option for VTT (maybe like .0325 higher or something). Could my board be damaged by not raising them?
thanks for your help thus far.

 

[SerpentRoyal]

To keep my E6320 stable. You can bump VTT, NB, and SB up one notch without adding much heat to the system. Higher voltage is often needed at high FSB speed for stability.

Your PC may not POST or pass Prime, but no harm will come to the MB with low voltage.

 

[monohouse]

something's strange about the screen shots at http://forums.abit-usa.com/sho...9267a7&t=128969&page=2
since when is a 4400 which is a cheep-crap version of the conroes has a multiplier as high as 10X ?

and the other strange thing is how could 3000 mhz be considered toasty for a conroe ? if 3200 can be achieved without any effort ?

how can memory speed be separated from NB speed on this board ? is there a divider lower than 1:1 ? like 0.80:1 ? I didn't see this in the bios

 

[mike525]

So should I run it at current settings first? Again, I currently have my fsb @ 475. My Vcore is 1.4525. I am running large fft's now. I will run this for a few hours then tighten memory timings and run memtest. If all is well, I will call it a day and deliver the pc to my brother-in-law. Does this sound OK?

 

[SerpentRoyal]

The best stability test is how well the PC run with your applications. There is no need to run Prime/Orthos for more than two hours. The main goal of stress test proggies is to heat up your system to cause failure. The PC may pass Orthos at 75F and fail at 85F.

Real world load is always 10 to 15C lower. If the PC is not stable with some programs, then back-off core speed by 50 to 100MHz.

If you change RAM timing, then you will need to redo Orthos.

 

[mike525]

Ok... Well, I have tried Battlefield 2, World in Conflict, and Fear and noticed no problems in any of them. I need to try Photoshop and that should do it. Is it even worth tightening the memory timings?

 

[SerpentRoyal]

The best stability test is how well the PC run with your applications. There is no need to run Prime/Orthos for more than two hours. The main goal of stress test proggies is to heat up your system to cause failure. The PC may pass Orthos at 75F and fail at 85F.

Real world load is always 10 to 15C lower. If the PC is not stable with some programs, then back-off core speed by 50 to 100MHz.

If you change RAM timing, then you will need to redo Orthos.

 

[monohouse]

I meen how can I run the ram at speeds that are lower than the FSB not higher

 

[hokiealumnus]

Originally posted by: monohouse
something's strange about the screen shots at http://forums.abit-usa.com/sho...9267a7&t=128969&page=2
since when is a 4400 which is a cheep-crap version of the conroes has a multiplier as high as 10X ?

and the other strange thing is how could 3000 mhz be considered toasty for a conroe ? if 3200 can be achieved without any effort ?

how can memory speed be separated from NB speed on this board ? is there a divider lower than 1:1 ? like 0.80:1 ? I didn't see this in the bios

Wish the abit forums weren't down (they're physically moving their headquarters in CA), but most likely with an E4400 over there, those are my screenshots. What issue do you have with them?

The E2xxx and E4xxx series both have high multipliers. E2410 is 8x, E2160 is 9x and E2180 is 10x. Similarly, E4300 is 9x, 4400 is 10x, 4500 is 11x and the new 4600 is 12x. Why are you angry about that? CPU-z reports it as a Conroe, but it's actually an Allendale core.

3.0GHz is toasty because most likely, in the SS you were looking at, I was using intel's stock cooler. With my Big Typhoon, running Orthos when my office is hot even it tops out at 63c. Now, it doesn't get toasty until I crank it up to see what I can do...say 3.4GHz @ 1.505Vcore is when it starts getting warm. 3.45GHz @ 1.565 still tops out below 80c (and yes, then I shut it down!).

RAM speed & NB speed gets separated by the ratios you reference. There is nothing less than a 1:1 ratio. If you're running a low multi chip and want to go >400MHz on your ram, you need DDR2-800 sticks that can overclock or DDR2-1066 sticks.

Some people have to build PCs on a budget, hence I went with an E4xxx chip. Why is it a crap chip? It does much more than what I was even expecting it to do frankly. Because it's inexpensive doesn't make it a bad chip. It's just one that works for some people on their budget. Oh, and cheap is spelled c-h-e-a-p, not cheep.

 

[monohouse]

yhe, and that is spelled "yes"
it's good to know, so what stucks you ? the temperature ? can you get to 4000 mhz ?
what is the difference between the allendale and the conroe ? conroe better ?

since when short questions mean that you are angry ?

im asking because if it's possible to get to 4000 mhz with a 4400 why did I buyed the mutch more expensive 6750....or is it really that mutch more expensive ?

do you think that if you had watercooling you could reach 4000 mhz ?
im guessing that the allendale is a lower-L2 sized version of the conroe...the 2MB not the 4, is that all that the 6750 has better than the allendale ?

it's a strange thing, my E6750 has a Tjunction of 100C and at max load using default voltage at 3200 mhz has an upper temperature of 47C with a room temp of 35C, that's about 12C higher, theoretically if the maximum temperature of the conroe is 75 for a 100C Tjunction I could increase alote more than that, like more than another 1000 mhz (originally I planned to buy the E6600 because of the 9x multiplier to get to 4500 mhz) but the E6600 was not available anymore so I had to go for the less-good E6750 and settle for 500x8 (4000) as a maximum, my intentions are to run the system at that speed at all times, but then I learned about the penryn (45 nm) and thought that maybe I can settle for 4000 mhz for the time being and then when penryn exists get a penryn with a 9x or 10x multiplier to get to 4500/5000 mhz

 

[hokiealumnus]

Originally posted by: monohouse
yhe, and that is spelled "yes"
it's good to know, so what stucks you ? the temperature ? can you get to 4000 mhz ?
what is the difference between the allendale and the conroe ? conroe better ?

since when short questions mean that you are angry ?

im asking because if it's possible to get to 4000 mhz with a 4400 why did I buyed the mutch more expensive 6750....or is it really that mutch more expensive ?

do you think that if you had watercooling you could reach 4000 mhz ?
im guessing that the allendale is a lower-L2 sized version of the conroe...the 2MB not the 4, is that all that the 6750 has better than the allendale ?

Sorry, I misinterpreted the "Since when..." part as anger. My bad.

What's holding me back I suppose are two things
1. Cooling; the BT is a wonderful air cooler, but you can't realistically run > 1.54Vcore on any air cooler, not just this one. Water, phase change, dry ice, etc would be necessary before my paranoid self would even attempt it.
2. My own aforementioned paranoia. Heat kills chips, yes but Vcore can kill them faster. Even running phase change, there is a limit to the voltage you can push before it blows up.

I don't think this chip could ever make it to 4GHz. 100% overclock would probably be asking too much. It may make it to ~3.7 if I pushed the hell out of it, but I don't have money to spend on a new chip.

Conroes are definitely better chips. I believe the major difference is the 2M cache on the Allendale vs. 4M on the Conroe. I believe there are other architectural differences as well, I just don't know them.

Your chip was worth the money you paid if you feel it was. You may be able to do what I will probably never be able to (unless I buy a new chip ) and join The C2D 4GHz Club! :Q

 

[monohouse]

my goals aren't about numbers, and I mean both numbers from benchmark results and numbers from getting a high mhz, none of them are the goals, they do however help me determine the performance of my system and make sure that it is running as good as it can, also I think "temporary screenshot overclocks" are crap because if you are not using your system at that speed at all times (like those phase change idiots that after the 1 minutes of running at 5000 mhz have to shut down before the entire system blows up) it's totally useless, the problem is that there is no way to be sure using a means such as the ones known to us for the moment that the person has actually executed heavy stuff on his system for long times, long enough to prove that his system is consistently (24h dual-prime95-affinity) stable, temporary overclocking is not good for anything, the only good overclock is the overclock that matters and it only matters if you use it .

and worse, the power requirements are several times that of the system itself, it doesn't make any sense to use sutch cooling because you can buy several computers after running this kind of cooling for so long

what do you think ? can a conroe be consistently running at 4000 without a problem ? for the moment let's assume with C1E/EIST even if it doesn't help that much

 

[hokiealumnus]

I could overclock my chip at crazy voltages and do some "temporary screenshot overclocks" if I felt like it. My goal was 3.5 for this chip. That's not my goal. I was looking for the maximum stable frequency I COULD run if I didn't care for the longevity of my CPU.

Serpent has said and my experience has confirmed that normal operation of a PC, regardless of application, will yield 10-15c lower temperatures than heating it up with Orthos/Prime. My goals for overclocking this were to see what would be a reasonable 24/7 speed.

The reason I mention longevity is because overvolting a CPU will eventually lead to instability. For instance, I have a friend that overclocked his E4300 and ran at 3.6GHz with 1.6Vcore on water for about a year. After that time, he started experiencing crashes and general instability. He tried raising Vcore a little bit and just couldn't get it stable, even up to 1.65v I believe. So, he dropped to 3.5GHz and backed back down to 1.6v. It became stable again at the same voltage but a lower overclock. The chip developed a 'tolerance' if you will. Another poster explained this better in another thread, but I like my analogy better.

My point is, I didn't go for temporary overclock screenshots. I wanted to see what I could reasonably run 24/7 if I wanted to. The overclocks I posted are perfectly stable, I just didn't want my chip to be over-volted 24/7 (until I can afford a new one, should it croak).

Anyway...to your question. Someone posted a topic over there about the longevity of the folks in the 4GHz club regarding how long they ran it. Only one person responded and he was running and E6850 at 4GHz (well, 499*8 for 3.99) at 1.58Vcore with a TR Ultra 120 for four months & counting. (Link)

You could run 4GHz for as long as your chip lets you. No one knows how long the process takes for it to need more voltage to remain stable. Well, I haven't seen it. I'm sure someone knows the properties of the materials and how they relate to the extra current.

To begin with, see what kind of voltage it takes for you to reach 4GHz and remain stable. Then see if your temperatures are reasonable. Then you'll have a decision to make. If it's stable at a low enough voltage that you think you have headroom to move when it decides it needs more current, go for it!

I'm talkative today...don't want to be sitting at this desk, that's for sure.

 

[monohouse]

what's reasonable voltage for a 100C Tjunction and G0 stepping conroe ? 1.6 ?

 

[hokiealumnus]

Originally posted by: monohouse
what's reasonable voltage for a 100C Tjunction and G0 stepping conroe ? 1.6 ?

That's the question isn't it? I have no idea. My personal limit was 1.565 (1.52 actual under load), and even that felt high for me. "Safe" is generally VID + 10-15%. My VID is 1.325, therefore my reasonable absolute max is 1.524 (where I stopped). You can go higher and certainly people have.

You can check out the Intel Processor Spec Finder to see what they recommend as your max voltage (1.5v in my case). If you don't feel like looking, it says the same for yours. (Link)

I wouldn't push past VID + 15% and probably wouldn't run that 24/7 (barring expendable income). Of course, this is all relative and you overvolt at your own risk.

EDIT - The IP35-E's default voltage is the VID for your chip, so it's the lowest voltage in that menu.

 

[goodthymes]

Quick question...

Recently purchased this Abit IP35-E, generally runs great but seems I've come across a curious problem. Previously I installed Vista on a Raptor 1st gen 36GB drive, no problem there - just that it's is loud.

Disconnected the Raptor, and installed Vista on a Seagate IDE drive. Installation completed without and problem. However, seems after installing the drivers for the motherboard and rebooted the boot partition is no longer detected? This is with only the IDE drive connected at the moment. Has anyone have any problem using IDE drives with this motherboard? Please advise, thanks much in advance.

The Seagate IDE is known to be good, and is detected in the BIOS, by the way.

E2140 (stock)
IP35-E
2GB Patritot DDR800 RAM
Evga 8800GT
Antec TP500

 

[jonmcc33]

Originally posted by: monohouse
so from 2660 mhz to 3200 mhz without even starting to mess with any voltages (but I will get to that later), it's a 540 mhz increase, there were times that (with my old Mobile 2500+ Athlon XP) I needed to fight to get that mutch, eventually I did (from 1800 to 2333 mhz) but the FSB of the mobo (KT333) was what stucked any further increases. and with this board more than 500 mhz without any efforts is just amazing.

my ONLY complaints about this board is: high memory latency (200 cycles) and double post during boot

I went from 2.13GHz to 3.2GHz myself. Over 1000MHz overclock...much better.

 

[Heidfirst]

BIOS 13 final released

http://www2.abit.com.tw/page/e...-E&pSOCKET_TYPE=LGA775

BIOS ID:13

Decrease the reboot times when adjusting BIOS.
Enhanced the IEEE1394 performance.
Add AC97 item under front Audio.
Support Wolfdale/Yorkfield Processor.
BIOS compiled date: 09/28/2007.

 

[harbin]

I found 32bit Server 2008 can use all 4x1G memory on IP35-E. Is this something special about this board? Or every P35 will do?

 

[SerpentRoyal]

Originally posted by: goodthymes
Quick question...

Recently purchased this Abit IP35-E, generally runs great but seems I've come across a curious problem. Previously I installed Vista on a Raptor 1st gen 36GB drive, no problem there - just that it's is loud.

Disconnected the Raptor, and installed Vista on a Seagate IDE drive. Installation completed without and problem. However, seems after installing the drivers for the motherboard and rebooted the boot partition is no longer detected? This is with only the IDE drive connected at the moment. Has anyone have any problem using IDE drives with this motherboard? Please advise, thanks much in advance.

The Seagate IDE is known to be good, and is detected in the BIOS, by the way.

E2140 (stock)
IP35-E
2GB Patritot DDR800 RAM
Evga 8800GT
Antec TP500

Could be old data in the HDD. If you have more than one primary partitions, then you may encounter similar anomaly. Save the old data on another HDD. Zero fill the HDD with the Seagate utility. Clean install windows.

 

[SerpentRoyal]

Originally posted by: Heidfirst
BIOS 13 final released

http://www2.abit.com.tw/page/e...-E&pSOCKET_TYPE=LGA775

BIOS ID:13

Decrease the reboot times when adjusting BIOS.
Enhanced the IEEE1394 performance.
Add AC97 item under front Audio.
Support Wolfdale/Yorkfield Processor.
BIOS compiled date: 09/28/2007.

Based on my test, this BIOS is the SAME as the BETA BIOS. No problem at all, except my 1M Super Pi time went from 16.250 to 16.610 with 13 BIOS.

 

[jonmcc33]

Holy crap. I have to use a floppy? Haven't had to use one of those for years...