Intel Pentium 4 2.40C

This time look a bit closer Intel Pentium 4 2.40C processor. As we know Intel introduced 3.0 GHz P4 along with i875P chipset and little later 2.4 GHz-, 2.6 GHz- and 2.8 GHz- models with i865PE chipset. So P4 processors now support 200 MHz FSB and hyper threading technology and what really kicks ass is that although 3.0 GHz model is quite expensive, 2.4 GHz is almost as cheap as "old" 533 Mhz 2.40 GHz. So 2.4C is obvious choice to our review as it support all new features is cheap and hopefully overclocks as madman. Right, this is mainly overclocking test as Internet is full of P4 reviews which quite often tell as same story in different words.

Processor:

I have to say that I cheated a little bit. I asked pretested processor from Akiba computer store here in Finland. So macci tested a few processors and send me a good one. Akiba actually sells 2.4@3.4 GHz 1.525 V (Air cooling, Zalman CNPS-7000CU) so there is quite a lot good processors floating around and if you don't want to take risk you can always buy pretested processor as I did. For a comparison I also tested quickly two other 2.4 GHz processors.

I'm not going to tell you boring marketing stuff here so if you actually want to know everything about a Pentium 4 2.4C take close look at this. However we should look few interesting specs mainly concerning our overclock attempt.

Processor

Voltage:

Pentium 4 processors now use voltages between 1.475-1.55 V. It seems that high-end models are using 1.55V and low end like our 2.4 GHz 1.525 V. I have not personally seen any 1.475 V processor but there might be some and it is quite possible that they are good for overclocking or at least they could be. And what is most interesting is that according to specs, absolute maximum voltage to processor should not exceed 1.75 V. This is idle voltage and on load it is normal that voltage decreases quite a bit. How much? Again I'm going to turn to datasheet and the truth is that if processors default voltage is 1.525 V, actual Vcore should be around 1.345-1.425 V. Suprised? At least I was when I first time actually read datasheets.

Temperatures:

As we know reducing processor temperature is one of the easiest way to get processor overclock like we want. But first we should look what the specs say about temperatures and cooling. Next list is partly (first three lines) concerning only 2.4C 1.525V processor at default clocks and voltages.

• Thermal design power (TDP): 66.2W
• Minimum temp: 5C
• Maximum temp: 74C
• Temp should ne measured at the geometric top center of the IHS (not the core temp!)
• Intel recommends that complete thermal solution designs target the TDP instead of the maximum processor power consumption.
• With a properly designed and charactized thermal solution, it is anticipated that the thermal control circuit (TCC) would only be activated for very short periods o time when running most power intensive applications.
• In the event of cooling failure the processor will automatically shut down when the silicon has reached a temperature of approximately 135C.

It looks like we should expect that 2.4C boxed heatsink is not very efficient. Specs clearly tell as that thermal solutions doesn't have to be at least sufficient to dissipate maximum power consumption because TCC takes care if processor is over heating. Good news are that boxed heatsink is actually quite efficient because TDP is much harder to dissipate under maximum case ambient temperature which is 42-45 C than under normal circumtances (20-30C). However I recommend better cooling solution for serious overclocking.

This is just a scratch of the whole thermal design thing and if you want to know more I recommend Intel's Thermal Design Guidelines Design Guide. So if we think logically I suppose that there is only few things that actually matters. First is that we should try to keep our test processor "alive". Second is that what lower the core temp the better the overclockability. Third is that the higher voltage helps and what lower the temperature that safer it is to use high voltages, even exceeding the absolute maximum ratings.

Possible problems:

2.4C uses 12x multiplier and because our processor is locked like every other P4, we have to make sure that our test machine tolerate as high FSB as possible. Quick calculation indicates that at 2.4@3.6 GHz FSB is around 300 MHz, which really is quite a lot. We have also another problem because I have not seen any memory module which can take 300 MHz, fortunatelly i865PE and i875P chipsets have memory multipliers like 5:4 and 3:2 so at the 300 MHz memory speed is going to be either 240 MHz or 200 MHz. Based on these expectations I choosed motherboard and memory modules for this overclocking test. Motherboard that I chooced is Asus P4P800 which offers us very good overclocking settings and memory modules are two 256MB HyperX PC3000 which easily can do 250 MHz 5-2-2-2.

Cooling:

Because I have decent watercooling setup installed on my case I decided to use only that. No aircooling this time folks. Watercooling setup consist Hydor Seltz L30 waterpump, WaterCube GT3 CPU-block, XCC Chipper I-Cu GPU-block and a big radiator. Power dissipation is not going to be problem because system is very well cabable of radiating much more heat energy than any computer is capable to produce. Within normal circumtances watertemp is only few degrees above ambient temperature with only few 26db 92x92x25 fans. Here is few pictures of cooling system.

Cooling

Cooling

Test processor 1#: Pentium 4 2.40C GHz:

Here is detailed information about a test processors.

• PROD. CODE: BX80532PG2400DSL6WF
• FPO#: L311A849-0736
• VERSION#: C30333-002
• PACKDATE: 05/12/03
• MALAYSIA

Test processor 2#: Pentium 4 2.40C GHz:

• PROD. CODE: BX80532PG2400DSL6WF
• FPO#: 3314A505-8479
• VERSION#: C30333-002
• PACKDATE: 05/30/03
• COSTA RICA

Test processor 3#: Pentium 4 2.40C GHz:

• PROD. CODE: BX80532PG2400DSL6WF
• FPO#: L310A700-0211
• VERSION#: C30333-002
• PACKDATE: 04/30/2003
• MALAYSIA

Nothing extraordinary here.

Processor

Overclocking processor 1#

I started OC-test by booting to Windows XP at default settings. Memory multiplier 1:1, MAM on, fastest memory settings and all voltages at default. Then I used SetFSB to raise FSB megahertz by megahertz. System booted nearly immediately and 12x210 MHz was best achievement with this method, not very impressive. So I set memory multiplier to 5:4, raised memory voltage to 2.85 V and booted to Windows (12x210 Mhz). This time OC went quite fine and system was completely stable to 3300 MHz (12x275 MHz). Still default vcore and I had feeling that processor could do much better.

Well CPU really was much better as I finally found best OC settings after countless crashes.

• CPU voltage: default (1.525 V)
• Memory voltage: 2.85 V
• AGP voltage: 1.8 V
• Memory multiplier: 3:2
• AGP/PCI frequency: 72/36 MHz
• Memory timings: 5-2-2-2
• FSB: 309 MHz
• Watertemp: 33 C

This is what I call impressive: 3710 MHz (12x309 MHz) at default voltage. Now my motherboard is seriously limiting my overclock efforts as 309 MHz is absolute maximum stable FSB. (I count system stable when I can run CPUBurnP6 and CPUBurnMMX as long as I want.) I'm speechless. Here is although few screenshots

CPU frequency

Memory frequency

Overclocking processors 2# & 3#

Well these weren't so good. Processor 2# could only do 3.2 GHz at default voltage and 3.3 GHz at 1.625 V. Processor 3# was little better and did 3.3 GHz at default voltage and little over 3.4 GHz at 1.625V. Very funny thing here is that processor 2# was supposed to be pretested (3.4 GHz 1.525V) but it obviously wasn't. Costa Rica chip seems to be looser here but I have heard that very new Costa Rica chips are actually quite good.

Tests

On this test sections we compare 2.40GHz processor against itself. I drove benchmark at default clocks and maximum overclock settings. Point here is to clarify how much overclocking can really boost overall system performance. If you want to know how much difference memory multipliers make you just have to wait Asus P4P800 review. On graphs there is number on bottom right corner which shows performance difference in percents.

WinRAR 3.10

CDex 1.50

PCMark2002

SiSoft Sandra 2003

Superpi 8M

Unreal Tournament 2003

3DMark2001SE

Quake III Arena

OC performance is just great. 25-50% better results for free is just something that I haven't seen since... ermm since my 1700+@2550 MHz. However performance is clearly superior compared to fastest available processors (AthlonXP 3200+ and Intel Pentium 4 3.2GHz). OK, perhaps I'm just lucky and processors like this are very rare but if you are looking for a new OC combo, P4 2.4CGHz and i865PE/i875P -motherboard might be the best choice.

Conclusion

Intel Pentium 4 2.40C   + It's cheap + Performance is impressive even at 2.4 GHz + OC performance just rocks - Perhaps too low multiplier - Not every 2.4C OC this much

Although on this review I didn't explained or tested such important features as HT or 200 MHz FSB it should be noted that they play very big role. For example if you overclock your 2 GHz celeron to 4 GHz it is still just damn slow compared to even to a 2.4 GHz Pentium 4. So basic features are the thing that makes processor fast and then OC is just something that can make it even faster. Results speak for themselves quite clearly, overclocking can make your cheap but still rich featured processor surpass most expensive and fastest processors out there.

Additional information