- Oct 13, 1999
- 22,377
- 7
- 81
Here is my mini-review of my experiences with the MSI PT880 Neo-LSR motherboard based on the new VIA PT880 chipset. For benchmarks, read any of the numerous reviews of the chipset at various hardware sites, such as here at Anandtech. My "review" will be different, really being more of my impressions of the board - something that I feel few published hardware reviews give. Sure, a review will show you what it looks like and how it performs, all the features and major flaws, but I want to know what is it like to actually own the part and use it and compare it with other parts I have.
I purchased this board from Newegg.com at around 1AM on Thursday, 3/18 for $59 plus $5 shipping. I received it around 3pm the next day (Friday, 3/19). Good job Newegg.
The board is the typical MSI red color. No super suprises for layout. Pseudo rounded floppy/IDE cables were a nice touch, as was the double SATA power adaptor. With so many included stuff wish they had included two IDE cables and two SATA cables (only one each).
I'm using a pair of classic Samsung 512MB PC2700, running at 1/1 ratio, default latency settings. CPU is a P4 2.53 pulled from a Dell purchased in December. The CPU has POSTed at 3.61GHz on an Abit BH7 with a healthy voltage boost, but didn't have a chance to find a stable speed since the Abit board had other issues... Other parts are an Abit Siluro GF4 Ti4200, Zalman CNPS-7000A-AlCu HSF, WD800JB HDD, HEC 300W power supply and a 48X CDROM pulled from a Compaq. All this is running on foam, without a case.
Results are that Windows XP installed without a hitch with the 2.53GHz CPU overclocked to 3.33GHz.
A Windows install can uncover instabilities. Also, I had run Memtest86 on a full pass with no errors (can also uncover instabilities). Haven't yet installed drivers or run any Windows-based benchmarks/tests yet (not caring about scores, but about running them without locking/BSOD/etc.). I decided to post this mini review for now, grab some dinner and play some Generals/Zero Hour online before getting back to this.
My road to those results was a rocky one. I started with upgrading my main machine's RAM. It was running the Samsung RAM at a 2/3 ratio and I wanted to put faster RAM in it. Oh yeah, it is an Abit IS7 board with a P4 2.6C overclocked to around 2.86 or so. I thought the RAM was holding back my OC, until I went with a Geil 1GB PC3200 dual channel kit. The RAM tested fine in other boards running at the rated 400MHz, CAS2 using Memtest86. If I try clocking my main system any higher, even with 2/3 ratio, Memtest would give errors. So, the RAM is underclocked at this time. Net result? 1 hour wasted, main machine running just like before except now can run CAS2 instead of 2.5. Hmmm, I had a 2.4C before I got the 2.6C and also couldn't run at high FSB, using various RAM up to PC3500 and trying all RAM multipliers of 1/1, 2/3, 4/5. Maybe it's the IS7 board... Anyways, now I had the Samsung RAM for the PT880. I figured that it would run 1/1 fine, giving me full dual channel bandwidth.
The BIOS of the PT880 has some irritations. On first POST there is a "Corecell" splash screen hiding the BIOS info. It could be disabled in BIOS. Then, the Southbridge RAID BIOS ran (can be disabled in BIOS). Finally can enter BIOS. Though those things can be disabled, they reset to enabled every time CMOS is cleared - and I had to clear it about 30 times!!! For a failed overclock attempt there is no automatic resetting of CPU settings like Asus boards have, and no "hold down Insert" like some other boards have. The clear CMOS jumper becomes your best friend and worst enemy. BIOS has some automatic overclocking thing, but I prefer to take matters into my own hands.
BIOS goodies: The PT880 does have a working PCI/AGP lock. VDIMM goes all the way to 3.3v and there are separate Northbridge/Southbridge/AGP voltage settings that I didn't mess with.
The MSI board has a maximum VCORE of 1.6v. PERIOD. End of story.
The HardOCP review mentioned the maximum voltage. BTW, I would rather have that versio of board, the PT880 Neo-FISR, because I actually DO use Firewire. Anyways, I figured that I could override the maximum using the "wire trick" to boost voltage. I was wrong. My CPU defaults to 1.525v and the BIOS can make it go 1.60v (in very teeny tiny increments). I shorted pins AE4 to AE5 on the CPU socket to make the default 1.55v. It worked. BIOS reported 1.55v, but still the maximum was 1.60v. I shorted pins AE1 to AE2 on the CPU socket to make the default voltage 1.725v (I had this CPU POSTing on the Abit BH7 at a higher voltage than this). No POST. Double, triple checked my work by taking out the wire and checking it at default, checking it at 1.55v, checking it at 1.725v. Going over 1.60v results in no POST. Why is that? This limits the board for hardcore overclocking of Northwoods. I think it is okay for overclocking Prescotts because they start at 1.35v (I think) and can be overvolted to 1.60v. Now, being that most of the reviewers couldn't get more than about 220-230FSB out of any PT880 chipset board, this makes the Prescott P4 2.4A chip the only shoe-in chip for this board, making for a nice performer and overclocker for those who have to own an Intel chip, but can't afford anything better. Hmmm, same conclusion that HardOCP got. I don't think this board would be suitable for massive overclocking of any other P4 CPU.
What is this board good for? Overclocking a P4 2.4A chip. Non-hardcore overclocking of any other chip. Non-overclockers. For people in those categories, this board is a very good choice, otherwise, look elsewhere.
For now I'll be content with getting my 2.53 to run at 3.33. I think this chip would do more, because it POSTs at 3.61 on an Abit BH7 using an older 845PE chipset that doesn't "officially" support any FSB higher than the chip default. I didn't try POSTing it higher on the BH7 because at that speed it wouldn't start booting off the hard drive - perhaps nearing the edge of the chipset stability? Thus, I think I'm limited on the MSI PT880 board with the 1.60v ceiling. Still, I'm getting much better results with this setup than with my Abit IS7 board, so it will be a keeper in a secondary machine.
I purchased this board from Newegg.com at around 1AM on Thursday, 3/18 for $59 plus $5 shipping. I received it around 3pm the next day (Friday, 3/19). Good job Newegg.
The board is the typical MSI red color. No super suprises for layout. Pseudo rounded floppy/IDE cables were a nice touch, as was the double SATA power adaptor. With so many included stuff wish they had included two IDE cables and two SATA cables (only one each).
I'm using a pair of classic Samsung 512MB PC2700, running at 1/1 ratio, default latency settings. CPU is a P4 2.53 pulled from a Dell purchased in December. The CPU has POSTed at 3.61GHz on an Abit BH7 with a healthy voltage boost, but didn't have a chance to find a stable speed since the Abit board had other issues... Other parts are an Abit Siluro GF4 Ti4200, Zalman CNPS-7000A-AlCu HSF, WD800JB HDD, HEC 300W power supply and a 48X CDROM pulled from a Compaq. All this is running on foam, without a case.
Results are that Windows XP installed without a hitch with the 2.53GHz CPU overclocked to 3.33GHz.
A Windows install can uncover instabilities. Also, I had run Memtest86 on a full pass with no errors (can also uncover instabilities). Haven't yet installed drivers or run any Windows-based benchmarks/tests yet (not caring about scores, but about running them without locking/BSOD/etc.). I decided to post this mini review for now, grab some dinner and play some Generals/Zero Hour online before getting back to this.
My road to those results was a rocky one. I started with upgrading my main machine's RAM. It was running the Samsung RAM at a 2/3 ratio and I wanted to put faster RAM in it. Oh yeah, it is an Abit IS7 board with a P4 2.6C overclocked to around 2.86 or so. I thought the RAM was holding back my OC, until I went with a Geil 1GB PC3200 dual channel kit. The RAM tested fine in other boards running at the rated 400MHz, CAS2 using Memtest86. If I try clocking my main system any higher, even with 2/3 ratio, Memtest would give errors. So, the RAM is underclocked at this time. Net result? 1 hour wasted, main machine running just like before except now can run CAS2 instead of 2.5. Hmmm, I had a 2.4C before I got the 2.6C and also couldn't run at high FSB, using various RAM up to PC3500 and trying all RAM multipliers of 1/1, 2/3, 4/5. Maybe it's the IS7 board... Anyways, now I had the Samsung RAM for the PT880. I figured that it would run 1/1 fine, giving me full dual channel bandwidth.
The BIOS of the PT880 has some irritations. On first POST there is a "Corecell" splash screen hiding the BIOS info. It could be disabled in BIOS. Then, the Southbridge RAID BIOS ran (can be disabled in BIOS). Finally can enter BIOS. Though those things can be disabled, they reset to enabled every time CMOS is cleared - and I had to clear it about 30 times!!! For a failed overclock attempt there is no automatic resetting of CPU settings like Asus boards have, and no "hold down Insert" like some other boards have. The clear CMOS jumper becomes your best friend and worst enemy. BIOS has some automatic overclocking thing, but I prefer to take matters into my own hands.
BIOS goodies: The PT880 does have a working PCI/AGP lock. VDIMM goes all the way to 3.3v and there are separate Northbridge/Southbridge/AGP voltage settings that I didn't mess with.
The MSI board has a maximum VCORE of 1.6v. PERIOD. End of story.
The HardOCP review mentioned the maximum voltage. BTW, I would rather have that versio of board, the PT880 Neo-FISR, because I actually DO use Firewire. Anyways, I figured that I could override the maximum using the "wire trick" to boost voltage. I was wrong. My CPU defaults to 1.525v and the BIOS can make it go 1.60v (in very teeny tiny increments). I shorted pins AE4 to AE5 on the CPU socket to make the default 1.55v. It worked. BIOS reported 1.55v, but still the maximum was 1.60v. I shorted pins AE1 to AE2 on the CPU socket to make the default voltage 1.725v (I had this CPU POSTing on the Abit BH7 at a higher voltage than this). No POST. Double, triple checked my work by taking out the wire and checking it at default, checking it at 1.55v, checking it at 1.725v. Going over 1.60v results in no POST. Why is that? This limits the board for hardcore overclocking of Northwoods. I think it is okay for overclocking Prescotts because they start at 1.35v (I think) and can be overvolted to 1.60v. Now, being that most of the reviewers couldn't get more than about 220-230FSB out of any PT880 chipset board, this makes the Prescott P4 2.4A chip the only shoe-in chip for this board, making for a nice performer and overclocker for those who have to own an Intel chip, but can't afford anything better. Hmmm, same conclusion that HardOCP got. I don't think this board would be suitable for massive overclocking of any other P4 CPU.
What is this board good for? Overclocking a P4 2.4A chip. Non-hardcore overclocking of any other chip. Non-overclockers. For people in those categories, this board is a very good choice, otherwise, look elsewhere.
For now I'll be content with getting my 2.53 to run at 3.33. I think this chip would do more, because it POSTs at 3.61 on an Abit BH7 using an older 845PE chipset that doesn't "officially" support any FSB higher than the chip default. I didn't try POSTing it higher on the BH7 because at that speed it wouldn't start booting off the hard drive - perhaps nearing the edge of the chipset stability? Thus, I think I'm limited on the MSI PT880 board with the 1.60v ceiling. Still, I'm getting much better results with this setup than with my Abit IS7 board, so it will be a keeper in a secondary machine.
Facebook
Twitter