Neat. Any big plans for it, or just going to hang onto it as a novelty?

And yes I would agree with your theory that the Radeon IGP is the problem - I vaguely remember the early ATi chipsets being derided for poor performance compared to Intel and even VIA counterparts. The Radeon 9000 itself, at least the desktop version, is a cut-down Radeon 8500, so it should make a decent DX8 machine if you're going to use it for gaming.

That's surprisingly slow. I've seen faster laptops with Mobility Radeon 9000 with lower clocked Pentium Ms. Maybe you can install newer (2006) Catalysts with that modded installer hack for better performance luck, or try forcing off PowerPlay

Oh wow, a netburst laptop.

You KNOW a netburst laptop just by lifting one up, they are the heaviest things in existence and have the most massive pure copper heatsinks you'll ever find. It's hilarious as low-end OEMs (acer) couldn't cut corners to reduce costs by using low-end cooling solutions since the laptops would melt down-- thus forcing them to use a lot of copper like the high-end OEMs.

3 Ghz? Well 3 Ghz with a massive pipeline and two discreet dies (which is unique now in the world of CPUs). So probably more akin to 1.5 Ghz.

EDIT: It uses the slower 533 FSB which is wounding-- it really needs an 800 Mhz FSB. Especially for prescotts.

leileilol wrote:

That's surprisingly slow. I've seen faster laptops with Mobility Radeon 9000 with lower clocked Pentium Ms. Maybe you can install newer (2006) Catalysts with that modded installer hack for better performance luck, or try forcing off PowerPlay

Weren't the Pentium M's supposed to be an improvement over the Pentium 4 based laptops? Better power usage, as fast or faster, etc. My router is actually a Pentium M laptop. 🤣 Pentium 4's were still during the megahertz/gigaherz wars

obobskivich wrote:

Neat. Any big plans for it, or just going to hang onto it as a novelty?

And yes I would agree with your theory that the Radeon IGP is the problem - I vaguely remember the early ATi chipsets being derided for poor performance compared to Intel and even VIA counterparts. The Radeon 9000 itself, at least the desktop version, is a cut-down Radeon 8500, so it should make a decent DX8 machine if you're going to use it for gaming.

No real plans for it. It's far too slow even for DX8 gaming. Heck, the GMA950 in my other old laptop (c2D, dual channel DDR2-667) is over twice as fast as the Radeon 9000 IGP. The single channel DDR-333, combined with the chipset's weak memory controller, just kills this thing.

I'll definitely hold on to it as a kind of collector's item. It's a laptop from a time when Intel had the nerve to call an 88w chip a mobile CPU. 🤣

leileilol wrote:

That's surprisingly slow. I've seen faster laptops with Mobility Radeon 9000 with lower clocked Pentium Ms. Maybe you can install newer (2006) Catalysts with that modded installer hack for better performance luck, or try forcing off PowerPlay

I'll have to give that a shot. Thanks.

raymangold wrote:

3 Ghz? Well 3 Ghz with a massive pipeline and two discreet dies (which is unique now in the world of CPUs). So probably more akin to 1.5 Ghz.

EDIT: It uses the slower 533 FSB which is wounding-- it really needs an 800 Mhz FSB. Especially for prescotts.

Well, my 1.59GHz PIII-S certainly doesn't feel much slower (if any) than this 3GHz laptop running the latest version of Firefox. Still, I believe ATI's memory controller is to blame here. Despite its architectural inefficiencies, Netburst at 3GHz should not be this slow. In fact, my Netburst "Reference System"--a P4 3.73GHz on an nForce4 Intel Edition board--is worlds better than this.

Truly mobile Netburst ended with the Northwood core with a final core speed of 2.5 ghz or so and a 400mhz fsb, everything p4 in a laptop from then on used standard desktop issue P4s and Celerons with the 533fsb northbridge being the only thing special. Overall the good ones people kept for a Long time as they didn't get too hot as they had pretty good cooling and really nice quality audio.

If you really want a Good Netburst laptop look to a Dell XPS or a Alienware/Sager/Clevo

Standard Def Steve wrote:

Well, my 1.59GHz PIII-S certainly doesn't feel much slower (if any) than this 3GHz laptop running the latest version of Firefox. Still, I believe ATI's memory controller is to blame here. Despite its architectural inefficiencies, Netburst at 3GHz should not be this slow. In fact, my Netburst "Reference System"--a P4 3.73GHz on an nForce4 Intel Edition board--is worlds better than this.

The 533 FSB is still pretty slow (although it may be due to the fact it's mobile). I have a 3.8 Ghz prescott, but it uses the 800 Mhz FSB which helps out a lot. It's by no means slow, but slower than a core2duo for sure. I keep it around as a workbench computer since the mobo has virtually every single interface ever made (from every type of SCSI, to SATA/IDE/floppy, to firewire/USB, etc etc etc). I also have XP on it as well, just in case there's a program that refuses to work in anything other than XP.

Wow, that is an awesome find! I rarely use laptops, but when I do, mobility and battery life hardly matter, I just use them as semi-portable computers I can use while sitting on a couch or in bed. It really is amazing you found a Prescott in that good condition.

I don't know what Toshiba P4s were like, but it might not be safe to run memtest86 for more than a short time on that laptop (the 30 seconds you showed is fine, I'd just not do a full length run). It might rely on Windows drivers to manage throttling so it doesn't overheat. Outside of Windows the cooling situation might be more dangerous.
I ran memtest86 on an HP dv6000 once - a consumer level Athlon64. The fan kept getting faster and faster and within a few minutes it got so hot the laptop abruptly turned itself off. It probably wasn't good for it.

shamino wrote:

Wow, that is an awesome find! I rarely use laptops, but when I do, mobility and battery life hardly matter, I just use them as semi-portable computers I can use while sitting on a couch or in bed. It really is amazing you found a Prescott in that good condition.

I don't know what Toshiba P4s were like, but it might not be safe to run memtest86 for more than a short time on that laptop (the 30 seconds you showed is fine, I'd just not do a full length run). It might rely on Windows drivers to manage throttling so it doesn't overheat. Outside of Windows the cooling situation might be more dangerous.
I ran memtest86 on an HP dv6000 once - a consumer level Athlon64. The fan kept getting faster and faster and within a few minutes it got so hot the laptop abruptly turned itself off. It probably wasn't good for it.

This is a very good point on more modern systems that have "adaptive" cooling. +1.

P4's throttle in hardware when overheating, Athlons don't. (Maybe this has changed for newer AMD CPUs.)

idspispopd wrote:

P4's throttle in hardware when overheating, Athlons don't. (Maybe this has changed for newer AMD CPUs.)

True, but by that time the CPU is pretty hot and I'd be worried about the thermal stress on the rest of the laptop.
In the case of that HP, it was notorious for heat related NVidia chipset failure, it wasn't really the CPU itself that was the problem, but throttling the CPU would indirectly help cool the chipset. When I had to replace the motherboard a couple years later, I resorted to a CPU undervolting utility just because of the influence it had over the chipset temperature.

shamino wrote:

Wow, that is an awesome find! I rarely use laptops, but when I do, mobility and battery life hardly matter, I just use them as semi-portable computers I can use while sitting on a couch or in bed. It really is amazing you found a Prescott in that good condition.

I don't know what Toshiba P4s were like, but it might not be safe to run memtest86 for more than a short time on that laptop (the 30 seconds you showed is fine, I'd just not do a full length run). It might rely on Windows drivers to manage throttling so it doesn't overheat. Outside of Windows the cooling situation might be more dangerous.
I ran memtest86 on an HP dv6000 once - a consumer level Athlon64. The fan kept getting faster and faster and within a few minutes it got so hot the laptop abruptly turned itself off. It probably wasn't good for it.

Well, the CPU is indeed throttling under load. It looks like CPU-Z just doesn't sense P4 throttling frequencies like it can with newer CPUs. Under 100% load, CPU-Z always reads 3.06GHz, no matter how heavily throttled the CPU is. Yet, it does report the various SpeedStep states properly.

I had it running memtest again today, and it overheated and turned itself off after around 25 minutes. The heatsink is spotless and the fans are both in perfect health, which means that it would've been impossible to achieve 3GHz performance for more than 10 or so minutes even when the machine was new!

In 3DMark03, the CPU tests begin around 8-10 minutes into the benchmark; by that point CPU performance is well below even an overclocked Tualatin! Although, the IGP may also be affecting the CPU score.

This machine (3.06/533 Prescott, Radeon 9000 IGP) = 293
PIII-S 1.59/151, GeForce 6800GT = 496
Pentium 4 2.4/400 Northwood, GeForce 6800GT = 558
Athlon XP 2800 (2.08/333 Barton), GeForce 6800GT = 679
Athlon 64 3700 (2.2GHz San Diego), GeForce 6800GT = 1257

Do Throttle Watch or HWiNFO32 tell you anything about thermal throttling?

Just adding a similar Vaio discard I've picked up.

P4 352, IGP 345M
512MB PC2100/DDR266
60GB HDD
15.4" 1280x800

As above, it relies on being throttled within XP. There are no BIOS CPU options. Memtest and suchlike are impossible. Installing Linux was also an ugly series of lockups. Finally got a Linux throttling app installed and the Vaio remained easy to push over the edge.

3DMark03 reported 'this system can only run 1 of the 4 game tests, try 3dmark01' and gave it 84 3dmarks on the default run. (Free version only does: cpu tests 0 if 2, feature tests 2 of 5, sound tests 0 of 3.)

3DMark2001SE scored 1295.

3DMark00 in 98 compatibility mode scored 2501.

My P3 450 MX440SE does 3172 in 3DMark00, and P4 3.06 7600GS does 10384 (also XP 98 compatibility mode).

It's dingo kidneys. That's a fresh install with drivers from Sony. Movies actually run slightly better on Linux for a change, but still not usable. Though perhaps that's only because the throttling isn't as aggressive.

The beautiful 15.4 display can only run low-res movies that you don't want to see so large and so clear.

There's some long old threads of frustrated people trying to get these ATI chips to play DVD movies. Found one period comment that out these Sonys were intended as budget take-home work computers; nice big screen for spreadsheets and side-by-side documents, and that's all. Seems to be true.

I'm kinda gobsmacked it could be so utterly useless for gaming, but the dependance on heavy throttling really leaves no options.

I had no idea. I'd heard P4 352 weren't all that good, but /wow/. This was an education.

pewpewpew wrote:

Just adding a similar Vaio discard I've picked up. […]
Show full quote

Just adding a similar Vaio discard I've picked up.

P4 352, IGP 345M
512MB PC2100/DDR266
60GB HDD
15.4" 1280x800

As above, it relies on being throttled within XP. There are no BIOS CPU options. Memtest and suchlike are impossible. Installing Linux was also an ugly series of lockups. Finally got a Linux throttling app installed and the Vaio remained easy to push over the edge.

3DMark03 reported 'this system can only run 1 of the 4 game tests, try 3dmark01' and gave it 84 3dmarks on the default run. (Free version only does: cpu tests 0 if 2, feature tests 2 of 5, sound tests 0 of 3.)

3DMark2001SE scored 1295.

3DMark00 in 98 compatibility mode scored 2501.

My P3 450 MX440SE does 3172 in 3DMark00, and P4 3.06 7600GS does 10384 (also XP 98 compatibility mode).

It's dingo kidneys. That's a fresh install with drivers from Sony. Movies actually run slightly better on Linux for a change, but still not usable. Though perhaps that's only because the throttling isn't as aggressive.

The beautiful 15.4 display can only run low-res movies that you don't want to see so large and so clear.

There's some long old threads of frustrated people trying to get these ATI chips to play DVD movies. Found one period comment that out these Sonys were intended as budget take-home work computers; nice big screen for spreadsheets and side-by-side documents, and that's all. Seems to be true.

I'm kinda gobsmacked it could be so utterly useless for gaming, but the dependance on heavy throttling really leaves no options.

I had no idea. I'd heard P4 352 weren't all that good, but /wow/. This was an education.

It could have been even worse as the Celeron D 352 is a Cedar Mill 0.65 um CPU, the version they used in some laptops (eventhough its not a laptop CPU) has an official TDP of "only" 65W!

Surprised the laptop lools so nice. Those mid-2000s Toshibas dont age well with amy use. You could probably put a lower clocked cpu in with negligible performance decrease for cooler operation. The 533mhz bus is probably the main bottleneck.

88w TDP on mobile CPU. Interesting, how large room it can warm.

I was checking the specs, and the 352 is largely identical to the Northwood HT 3.06 in my XP gamebox. Differences are 1024kb L2 instead of 512, and 88W TDP instead of the desktop's 82...

In winter, I do enjoy how the Northwood takes the chill off the room. It's taken up the slack left from replacing the CRT with LCD.

idspispopd wrote:

Do Throttle Watch or HWiNFO32 tell you anything about thermal throttling?

Must say I don't understand how this is working.

Wikipedia: "Under Microsoft Windows XP, SpeedStep support is built into the power management console under the control panel. In Windows XP a user can regulate processor speed indirectly by changing power schemes. The "Home/Office Desk" setting disables SpeedStep, the "Portable/Laptop" power scheme enables SpeedStep, and the "Max Battery" uses SpeedStep to slow the processor to minimal power levels as the battery weakens. The SpeedStep settings for power schemes, either built-in or custom, cannot be modified from the control panel's GUI, but can be modified using the POWERCFG.EXE command-line utility."

I'm on AC power, in Home/Office setting. So what throttles it to 1860 when it's under load like this? FWIW it went back into the 2000s immediately when VLC crashed. Slowly climbing till it BSOD'd while I was trying to get it to allow USB removal.

Did try HWINFO32. Get BSOD when I try to run it, and when I try to uninstall it...