p4 vs core duo 2

i have a desktop p4 2.8ghz with hyperthreading and i have a laptop core duo 2 1.5ghz. i have performed benckmarks in which the core duo won significantly. but it fail to run games as blacksite area51 and i believe it is the ghz in the core duo.i have not tried the p4 because the video card is crap (nvidia 5200)

my question is should i continue to use my laptop for gaming until i can afford the quad core, 8800 sli notebooks. or should i put a 8600gt in the desktop until i buy that dream notebook.

btw the desktop is only agp and regular pci. not pci 16 express

 

The 1.8GHz pentium will bring down framerates by a lot. Hpyothethically, even if you bought a 8800 Ultra it wouldn't perform too well with the pentium 4. BTW, I have the same Pentium 4 2.8GHz processor as you, though I'm not entirely sure of it's capable of multithreading.

Your notebook should suit you better than your desktop for Blacksite.

 

it has hyperthreading depending on the model u have a easy way you can tell is if your l2 cache is 512kb then you have hyperthreading and if it is 1mb then you dont. and you have to set it up in your bios.

but in blacksite i get 10-12 fps and i see the cpu is both at 100% and i know it is not using all 1.5ghz effeicently. do you think the p4 is that bad

btw i know my videocard is bad on my laptop

 

A 2.8ghz p4 is still much slower than a 1.5ghz core 2 duo because of the architecture as well as intel's newer ghz scale. The 1.5ghz actually translates to around 2.5ghz or something like that in the P4 scale. Your laptop graphics card isn't that bad, just turn down some of the graphics settings and resolution in the game to improve framerates.

 

there is no real settings in that game not to get off topic



what do you mean buy intels new ghz scale i havent heard of this before

 

You should not be having trouble running Area 51. Here is a link to check the minimum and recommended hardware. Both your desktop and notebook should run it.

 

Basically intel doesn't use the same ghz to rate the speed of their processors. I'm not exactly sure how or why they changed it, but generally anything core duo or newer x about 1.5 or 1.6 = equivalent p4 speed.

 

ive tried that website it does not have blacksite area 51 only old area 51

 

I don't know why your games aren't running the way you like, but on all counts your laptop should kill that old desktop. Even a measly T5250 should kill the P4 and the 8600M GS should kill the FX5200 (which is, to put it bluntly, a piece of junk. I know because I have one. In my server).

I reckon you've just got crummy drivers.

 

They do use the same Ghz to rate CPU's, a Ghz is a Ghz. What has changed is the amount of work accomplished per clock cycle (Ghz). To explain, you have two pick up trucks, 1 is a 1/2 ton the other is a 1 ton. They are both fully loaded with bananas. They go to market leaving from the same place traveling 60mph (Ghz) who is going who will deliver more bananas upon arrival having traveled the same speed? Same with different CPU family's.

 

yeah i understand a ghz does not change in value, just like a meter will always be a meter, but there is a multiplier you can use to calculate approximately the speed of a core duo or later processor in terms of p4 standards.

 

i know my 5200 is garbage that is y i wondered if i should get a new video card but since my laptop outperforms it i guess ill stay with for now

and my drivers are up to date i go to laptopstogo and get it

 

my core duo has 9x
my p4 has 14x
checked by pc wizard
is the lower the better

 

No not what he is saying.

 

The multiplier times the front side bus (FSB) gives you the clock frequency, so no; the multiplier is no measure of performance.

In fact, since the Pentium M (and possibly before) clock speeds have no longer mattered; when the Pentium M came out for example, the 1.6GHz model could match the high-end 3GHz Pentium 4 Prescott model. Its all about architecture, and most recently, its also about how many cores you have.

 

im still amazed at the fact a 1.5 dual core kills a 2.8 w\htt

 

Apart from architecture and number of cores, the manufacturing process also matters; a Pentium 4 Prescott has transistors 90nm in size while a Pentium M has a smaller transistor size. The Pentium M can then fit more transistors on a die the same size, and thus increase processing power.

 

Be clear that, that is in multi-threaded benchmarks! Two CPU's. You can first off double on those as you have two, then add the increased efficiency! Not a real world comparison! By the size of the cache? NOT! A 1.5 C2D does not beat faster P4's in many things. And jumping to a quad core is the worst idea. I mean if dual is not supported quad will not be either!

 

Your 1.5 GHz T5250 is almost exactly equal in power per core to the 2.8 GHz Pentium 4 (1.85 being approximately the multiplier in purely CPU intensive applications, and Dynamic Accleration affecting that slightly). It's about twice as powerful in mutlithreaded benchmarks, but that isn't going to help you with single-threaded applications.

So switching to your desktop, even with a good video card, wouldn't really help unless you upgraded it's processor. It's either 8600 GT + 3.x GHz P4 in the desktop or a better processor in your laptop, probably the cheaper of the two options. Or settle for the performance you're getting now.

Before upgrading the Pentium, be sure of which model it is and which socket it uses. Socket 478 limits you to certain processors up to 3.4 GHz, while LGA775 allows up to a full 3.8 GHz. It's important to get the right type so it fits on your computer.

 

At this point I think your 1.85 multiplier is ****ing nuts per core! Where do you pull this from? I seriously doubt and if you post back more I will express more of what I mean.

Edit: I do want to hear, I do want you to explain.

 

Guys,

Keep the discussion civil and please do not bypass the language filter.

Thx
Kdawgca

 

It's an approximate comparison, based mostly on my own observations, as Pentium 4 benchmarks aren't near the top of the Internet waste pile nowadays. But the 1.85 multiplier has been more or less consistent in all the tests I've done:

1. Game comparison - Civilization III. Playing the same AI turn on both a 2.66 GHz P4 Northwood and a 2.2 GHz Core 2 Duo. AI turn time was 46% lower on the Core 2 Duo. 2.2 divided by .54 is 4.07, which is nearly exactly 1.85 times 2.2. There was slight error due to the necessity of clicking buttons as quickly as possible, but given the 10-minute AI turn time on the Pentium 4, it's still in the 1.8 to 1.9 range.

2. Comparison using my own Java program (single-threaded). The program calculated determinants of matrices using cofactor expansion, and kept track of the time it took itself (thus resulting in less human error). This time I compared a 2.2 GHz Core 2 Duo to a 3.2 GHz P4 with HT. Extropolating the data, the Core 2 Duo was equal to a 4.00 GHz P4 with HT (presumably Prescott). This gives a multiplier of 1.82. Prescott may be slightly more efficient per MHz, but this is still in the 1.8 to 1.9 range.

3. Super Pi. Look at the Bye Bye Super Pi thread. Results are scant, but an easy comparison is my 2.66 GHz P4 in 23rd place for single-core vs. #12-#14 in the dual-core category, all 2.66 GHz Core 2 Duos. Remembering to double the dual-core times, you get 67.2 seconds for the Core 2 vs. 119.8 for the P4 Northwood. The Pentium 4 takes 1.78 times as long as the Core 2 Duo of the same speed.

So perhaps 1.8 would be a better estimater; the 1.85 test was the first one I did and I've been using it subsequently as a consequence. However, the lower the Core 2 Duo (and thus the higher the affect of Dynamic Acceleration, the higher the Core-->P4 multiplier goes).

My poor 1.4 GHz Core 2 Duo owning friend will be most disappointed to hear his P4 equivalent is now 2.52 instead of 2.59 before Dynamic Acceleration re-adjustment.

 

OK let me get this correct, you think a 1.5 C2D is X1.85= 2.775 X2= 5.55, I simply have doubts. You either miss state or misunderstand. There is no way a 1.5 C2d is equal to a 5.5 P4. Also there are so many P4's you can not say. Redo your math and repost!

Edit: Meant in the best spirits, but think your 1.85 is way too high?

 

im lost i can brnckmark my two cpu and show the scores

 

I'm sure I posted this previously: Some results for running the same mathematical modelling scenario on some different hardware (tests done in Jan-Feb 07):

• Intel T7200 (2GHz Core 2 Duo) with 2GB RAM ....... 2.9 hours
• AMD Athlon 64 3200+ (2GHz) with 1GB RAM .......... 3.4 hours
• Intel Pentium 4 2.8GHz with 1GB RAM ............ 4.5 hours
• Mobile Pentium 4 1.5GHz with 256MB RAM .......... 8.6 hours

This shows the T7200 being around 50% faster than a 2.8GHz P4.

The model was single-threaded so, in reality, the T7200 could (and I sometimes did) run two models at the same time with very little hit on the time taken to run one so overall it could run three (or two plus two halves) in the time it the 2.8GHz P4 to complete one run. I didn't see strong evidence of the CPU cache size being a significant factor in this test.

John

 

did that p4 have hyper threading on

 

Do you have the multicore hotfix installed on the laptop? Also you could try using imagecfg to force the game to run on one core. I had issues with some games and dual core on a C2D laptop and those two fixes sorted it out.

 

I honestly think the GHz P4 scale is baloney. I don't think a P4 (even if it could scale to 5.5GHz) scales linearly so you can't just multiply the processing power of a certain speed of P4 and say so and so chip is equal to this many GHz of P4. And of course there are different P4 cores so you'd have to specify, blah blah blah.

Just my two cents.

 

i think i did that installation on my vista x64 but not xp x86 and how would i reduce it to one without using cntl,shift,esc

 

I haven't a clue. It was a 3+ year old computer in the office.

John

 

we all forget sometimes

 

More a case of never known, although my brain has become less retentive with age.

John

 

what do u think about the pentuim d my cousin has a 2.8 pentuim and my 1.5 c2d has the same windows index score of 4.6.

my question is did inyel improve the pentium d or is it crap. because its cheap

 

Upper end Pentium D only trails a lower end C2D by a bit (specifically try comparing the Pentium D 960 and the E4300).

http://www23.tomshardware.com/cpu_2007.html

That Pentium D will suck up twice the power for similar performance, though.

 

Is that the overall WEI, or for the CPU? I presume the latter (but WEI is not a recommended benchmark - I've observed a CPU score of 5.1 for a T2250 and 4.9 for a T7300).

As already noted, the Pentium 4 series best feature was as room heaters during the winter. AMD were quite willing for Intel to steal the heat crown. The Core / Core 2 series are a development of the Pentium M and are much more power-efficient.

John

 

it was the cpu score and i know u answered this but the p4 w\htt 2.8ghz scored 4.1

 

But WEI is a synthetic benchmark and generally a pretty poor one at that. It doesn't really give too great an idea of "real world performance".

Even so, if you want to compare synthetic benchmarks you're better off looking at a 3DMark06 CPU score or something like that.

 

is there any other programs cause his graphic caed is intel based

 

PCMark05 may be appropriate and there are plenty of results in the various notebook reviews.

John