Northwood 1.6a, any OC difference between Malay or Costa Rica? Updated-Some Early Conclusions

[vspt]

cyberizer9,

sorry didn't reply to your post about the maxtor drives, since I'm stuck at home using walmart connect which is the same ip's as aol's, so I can't post to the forums. at work now, but anyways, got my drives at staples.com, used a 20 off 100 and the free computer bag coupons, needed something for my labtop.

vspt

 

[Kroffty]

WOW I just booted into windows XP pro and got 1/2 way through 3dmark2001se and it crashed.
running at 2592mhz p4 1.6a costa rica @ 1.75v
Thats 162mhz fsb!!!! on crucial 2100ddr mt chips
p4b266c
Its totaly stable at 2500mhz 1.7v prime95 and sandra burn in test.

now I'm wondering if I need more voltage or better memory?

 

[SexyK]

Anyone have any experiance with Mushkin RDRAM, ie what speeds it can handel? It's got samsung chips on a custom PCB i believe. Anyone having any luck with this stuff?

Kramer

 

[Flatbroke]

I got 512 Mushkin 2-3-3 (133)from Newegg. It has Nanya chips. I looked up the part number on the chips and saw that it ran 143 at 2.5 cas. I got my CR 1.6A up to 140, but only got to try 3dmark2001 a couple times before I ran into trouble with my video card. That was at 1.625 volts. Will do more testing tomorrow when I start troubleshooting video card.
Edit: just saw question was about RDRAM, I have DDRAM

 

[WheelsCSM]

I just finished 5 hrs of Prime95 on my Costa Rica at 2.4 GHz (150 FSB) I got it to boot at 2.5 GHZ, but wasn't stable. I could probably do it if I wanted to raise the voltage, but I'm happy here (1.5 Volts), besides, my temp under load is 49 C. What is the max these can take? BTW, I'm using a P4S33 with 256 MB of Crucial PC2100.

 

[PlatinumGold]

P4 1.6a, Malay chip. ordered it from Newegg last week monday got it today. holidays suck for that stuff.

anyway, preliminary results.

FSB Ratio Total Speed Result Ram Mhz
133 1:1 2128 Runs 266
138 1:1 2208 Runs 276
140 1:1 2240 fails to boot 280


140 4:3 2240 Runs 210
145 4:3 2320 fails to boot 217.5

my suspicion is that right now, it's not the ram that's holding me back. if i up the voltage i can probably get better oc's, but i'm going to refrain.

 

[Slacker]

I received my 1.6a from googlegear today, it is a malaysia chip, using it right now at 145fsb 2320mhz it wouldnt post at 150 even at 1.7volts , but that may be the motherboard, it is running around 46c at 1.65v , 4114 in 3dmark2001 up from 3500 , very nice scores in sandra, beats all of the comparison systems except the 850chip mem bandwidth test, I have 2175mbps mem bandwidth 850/ddr gets 2450.

I have run 3dmark 2001 many times, and played some heavy Q3, both of those programs are good at crashing unstable overclocks, everything is looking good so far I would have liked to get 150mhz cpu fsb for 2.4ghz and memory at 150 for 300 ddr (I have DDR333 memory) but I am not going to worry about 80mhz on the cpu and 43mhz on the memory.

Is the 1.6a the new DANCING QUEEN I say YES!!

 

[striker21]

<< edit: I forgot to mention the one positive change that I am noticing is how quickly the temps change in Asus probe under full load. With the thermal pad it would take quite a bit more time as compared to AS3 to reach top temp. Also, once CPU activity ceases the cpu temp returns to idle temp much faster using the AS3. >>



Isn't this a bad thing for AS3? CPU temps are measured from the core with the P4, right? If the CPU temp is rising faster under load with the AS3, doesn't that mean it's doing a worse job of removing the heat than the thermal pad? It doesn't seem like the goal would be to make your temperature delta "more responsive"...

 

[Flatbroke]

Sounds like it to me too. I also removed the pad and used ASII, it was because of reading that after the stock pad melted, you have to lap the heat sink to get the surface clean to allow the AS to fill the microscopic scratches. (per AS instruction page). At idle running 16x140, with 3 34cfm fans, case is 26c and chip is 33c. After 3 hours of Prime95, (stable enuf for me, no crashes there or in games), chip is 49c. I am beginning to think that the design of the stock pad is getting better from reading some of the before/after reports. We need some tech site to test this, (hint, hint)

 

[prontospyder]

I got my Asus P4B266-C and P4 1.6A Costa Rica chip from Newegg on Tuesday. It runs fine at 2.133 Ghz with these settings:

133Mhz FSB
1:1 Memory
1.5 Volts
Turbo 1 mode

Is there any other setting that I'm missing? Where can I check to see whether it's running at 66Mhz AGP/33Mhz PCI? How come in Motherboard Monitor 5 it shows the voltage but the temp is 0 degrees C? Thanks.

 

[Errtu]

Is it just me or does it seem that alot of people are getting successful 1.6A OCs running at default voltage? (1.5v) Many are doing 133MHz FSB no problem, and I've seen a couple posts of people doing 150MHz FSB...all at 1.5v!?!?

 

[DAPUNISHER]

1.6A Malaysia/batch#L147B217 and P4B333/ from Mwave.com (FED EX pooched on my Samsung 2700 from Newegg so it won't be in till later today) I'm running the 133/133 1.1 combo@ 2.133ghz default voltage using Crucial 2100DDR. I'll push it harder once it's been running a week or so stable at this voltage/clock speed with the 2700DDR in there. It looks good in benchies but it's way to early to call it "rock solid" so only time will tell.

 

[iahk]

Yep, I'm doing 2133 on the 133mhz fsb at stock voltage with no prob. I'm even using temporary crap PNY DDR2100 memory from best buy. I'm still waiting for my damn kingmax ram to arrive. Should be here today though.. hmmmmmmmmmm...

BTW- Costa Rica chip

 

[dowxp]

http://forums.anandtech.com/messageview.cfm?catid=27&threadid=720573

P4 1.6A from googlegear
packdate of 1/16/2002
batch L147B212
Malay

http://www.ghg.net/shukuo/p4/

preliminary 1.6 @ 2.603, ran fine in windows.

http://www.ghg.net/shukuo/p4/final.avi - movie of final FSB; 16mb

random reboots in rc5/seti. had to back down to 160fsb. If i had better cooling solutions, perhaps i could run more stably at 163 or perhaps go higher. i have no gone higher that 163. running at 1.65V

 

[manko]

<< I also removed the pad and used ASII, it was because of reading that after the stock pad melted, you have to lap the heat sink to get the surface clean to allow the AS to fill the microscopic scratches. (per AS instruction page)...I am beginning to think that the design of the stock pad is getting better from reading some of the before/after reports. >>

My chip hasn't come in yet, but are you suggesting to leave the thermal pad on and use AS?

 

[Flatbroke]

removed the pad and used ASII

Not at all, I removed the pad and used AS exclusively. Pad is just fine for 133fsb IMO, if want to go higher or just want to spend 1/2 to 3/4 hour cleaning chip and stock HSF to put on AS, go ahead, although some people are reporting better results with the pad. (In some cases I think mistakes were made in the proper application of AS)

 

[dowxp]

<< removed the pad and used ASII Not at all, I removed the pad and used AS exclusively. Pad is just fine for 133fsb IMO, if want to go higher or just want to spend 1/2 to 3/4 hour cleaning chip and stock HSF to put on AS, go ahead, although some people are reporting better results with the pad. (In some cases I think mistakes were made in the proper application of AS) >>



dunno. 1.6 @ 2.608 using the pad and ORIGINAL AS. i havent used up the tube from 2 years ago yet.. >_< 40/50 temps

 

[Flatbroke]

Where did you read about using both? This seems very wrong to me.

 

[SexyK]

I got my newegg Costa Rica 1.6 yesterday. I've got 384MB Mushkin PC800 on a TH7-II-RAID. At first, it wasn't so stable at 2.133 @ 1.5v, so i ran it there for a few hours at 1.55v. Then when I bumped it back down to 1.5v everything was running beautifully. Haven't tried to push it past that yet. I probably will, but it's damn fast where it is!

Kramer

 

[manko]

I'm holding off on a motherboard for a few more weeks to see what comes out, but I'm hoping to get a rock stable 2.4Ghz from a 1.6a with the stock fan (more because it's quiet rather than cost).

I already ordered some AS3, but you guys think I should just go with the thermal pad and forget about the AS3, since I'm not planning to OC over 2.4? Is it a real pain to get the thermal pad off after burning in the chip?

 

[Errtu]

CPU: 1.6A @ 2.32GHz
Voltage: 1.5v (auto)
Cooler: stock Intel HSF w/AS3
Fpo/Batch: 3149A726
Location: Costa Rica
Pack Date: 1/25/2002
Motherboard: Gigabyte GA-8IRXP
FSB: 145MHz (580MHz quad-pumped)
Memory: 512MB Corsair XMS 2400 (PC 2400, CAS 2) clocked at 290MHz (2:1 ratio)
PCI/AGP Frequency.: 36MHz/72MHz

System posts at 150MHz FSB (2.4GHz) but bluescreens windows on boot (default voltage). Will try cranking up the voltage, but I know if I set it manually I won't be able to reboot (only problem with this board).

I've been running prime95 for about an hour now with no hiccups at 2.32GHz. Temps idle are 36c...haven't seen anything over 49c yet.

So far I'm dazzled beyond belief! Thank you Intel and Newegg!

[Edit:] Boots Windows at 150MHz FSB @ 1.65v... 2.4GHz so far seemingly stable! Temps have gone up a few degrees though (up to 53c so far) while running Prime95 for only 15 minutes. I'll probably go back down to 145MHz FSB just for the lower temps.

 

[DRS]

<<

<< edit: I forgot to mention the one positive change that I am noticing is how quickly the temps change in Asus probe under full load. With the thermal pad it would take quite a bit more time as compared to AS3 to reach top temp. Also, once CPU activity ceases the cpu temp returns to idle temp much faster using the AS3. >>



Isn't this a bad thing for AS3? CPU temps are measured from the core with the P4, right? If the CPU temp is rising faster under load with the AS3, doesn't that mean it's doing a worse job of removing the heat than the thermal pad? It doesn't seem like the goal would be to make your temperature delta "more responsive"... >>




If the temp is indeed measured from the core then yes that would be a bad thing. ( I just unplugged the fan to the heatsink and MBM5 is still reporting a temp that fluctuates with cpu activity so it certainly seems that way ) Also, from flatbroke's and my experience it looks like you are better off with the stock thermal pad!!!

We could run one last test to really be sure. I'm noticing that while running P95 torture, with the fan unplugged, the temps are relatively stable (i.e. they have not shot up quickly). In fact, after running P95 torture for ~5 minutes I just now hit 50C. Unfortunately, temps are not lowering after I stop P95 torture (of course they are not rising either but I'm willing to bet they would if I started P95 again). So, it would be interesting to make a comparison against the stock thermal pad under the above conditions.

BTW It's just hovering at 51C with little to no CPU activity after another 10 minutes.

(ABTW, there's something satisfying about burying MBM5's Mhz needle )

 

[Flatbroke]

Just to be clear, I removed the pad from the heat sink. I never ran it with the pad. I installed the heatsink with ASII and get 36-37c at idle and 49c after 3hrs of Prime95. Temps rise quickly and drop quickly when load ends. Temps should rise slowly and drop quickly, but I am interested in max temps under load, not how fast they rise. I have no experience with the stock pad. It was by design that I went right to the AS, to keep from having to remove metal from the heatsink, (per AS instructions), if I used the pad first and then wanted to use AS. Again, we need a test of stock pad vs. ASII by a tech web site.

 

[SubfornicalOrgan]

Retail Intel Northwood 1.6A
$150 from NewEgg.com
Costa Rica, SL668, 01/31/02

Retail Asus P4B266/PA/WOL-UAY
$80 from CDW.com

Rock solid 16 x 150 = 2.4GHz @ 1.65V and 40 deg C (idle) w/ retail heatsink and thermal pad

Note: The onboard audio didn't like it when I ran the FSB @ 150MHz so I had to disable the onboard controller. Anyone know of a good deal on a pci sound card? j/k

I haven't had time yet to try lowering the voltage or pushing the FSB further but I think it has a little give left in it.

wcpuid screenshot
SiSoft Sandra screenshot

...For reference, the following is a SiSoft Sandra screenshot of my 1.6+ XP @ 1.5GHz (1.8+):
SiSoft Sandra screenshot

The Northwood wins this comparison but both are good chips.

 

[striker21]

Flatbroke: I agree. Seems like slow-rise, fast-drop would be best. Yes, it would be nice if someone could compare the effectiveness of the thermal pad with AS3. Based on some comments from Nevin in his interview at Van's Hardware, it seems like the integrated heat spreader of the P4 would make the difference in thermal interface material less important than for something like an Athlon. The reason is the significantly increased surface area of the P4's heat spreader vs. the Athlon's exposed core (which is relatively tiny).

Anyway, I would also like to see someone take a scientific approach to this and try to make heads or tails of it. I might fork a thread on this topic, since it really doesn't belong in this one...