Level 3 cache detected?

Hi. I was running a test and i found there was a level 3 cache. Normaly, like CPU-Z and other softwares detect only 2 cache. Whats wrong? Here a screenshot.

I done the test like 10 times.

I believe the question to that answer is pipelining. It's very easy to fit <512KB of data into the cache (which is 512KB) without having to refer to system RAM more than once. If you'll notice, his latency takes a huge jump when going from 512KB to 1024KB. The processor has to request the data from ram twice (however many cycles that is I dunno) to fit it in the L2 cache. Why the latency remains relatively equal with 32MB of data, I have no clue. According to my logic latency should increase exponentially with an exponential increase to data size. So maybe I'm wrong.

I think that's the answer. Does that make any sense?

-mpjesse

not to me

OH. thought you were using the northwood version of the 2Ghz P4.

err.... did intel even make a 2Ghz northwood?

Anyways, nevermind my post then.

hell if I know. LOL. I would trust CPU-Z's cache reports more than that benchmarking program.

What does CPU-Z say?

-mpjesse

woodies got down to 1.6ghz? wow. never knew that...

Here CPU-Z screenshot.

What do you mean?

There's no such thing as an L-3 cache in a processor. L-3 cache can only be found on the mobos, but not all mobos do have L-3 cache.

I agree I have a Laptop with the 3.4 extrem edtion with 2mb L3 cache.

Anyway i buyed a new computer. I will recieve it tromorow. Here the specs:

Pentium D 920 2.8 Ghz
Unknown motherboard
GeForce 6200SE 256mb with Turbo Cache
Windows XP Media edition
With a remote controler.
2x 1028 MB Ram
250 GB 7200 RPM HDD

Is that a good choice? I hear the Intel was hots.

it's OK... but the video card is complete garbage. I'm not sure how much it cost but a 6600 is minum for gaming now. Also, that intel dual core is low end i think, proably should have got an AMD system with 3800+ x2 since AMD has better, cheaper, cooler dual cores.

Also, you seem to always rely on just numbers to decide which component is better. That 6200 may be 256mb but it sucks ass, while a 6600gt with 128mb will own it completely. Same for the RAM and Hard drive, the ram may be 2gigs, but is it crappy kingston value RAM or Corsair xms, not specifically, but thats a range. The hd might be 250gig but could be really slow and hot, but don't worry about that so much. Try to find out which mobo you got, it will help.

It sounds like you went and got an MDG package, and they may be fairly cheap, but they are crappy, how much did you pay?

Lohl, that's nothing
New Celeron Cedar Mill's hit 5.0ghz on stock intel cooling

yeah, but that's phucking LN2. This is air cooled..Intel cooled.

Lawl, thats basically the 3.33 to 5ghz

Why do you need to find that anyway?

Random red-flag: unless I'm mistaken, you appear to have two sticks of RAM with different timings. This is generally a bad idea.

My computer is not home builded. I buyed it. Here the full computer name:

HP Media Center PC m7350n

And for my old computer, i had only 1 stick of RAM. 512 mb.

Why do you need information on your motherboard? I'd be glad to show you some things if you just tell me why...

I wonder if it is related to hyper-threading?

d00d, you got a cheap system, just be happy it works.

Help how? Everyone knows that HP specifies their own boards, and that mid-range HP's generally use Asus boards that are modified from the standard version by removing features.

You have 2 pages of replies over a 2-day period.

I would suggest not insulting the people that are trying to help.

If you don't get an answer to your question, please rephrase it or be more specific. Your english is a little choppy and people might not understand the question.

Well first, your BIOS is out of date. Try updating that first.

While CPU-Z reports your BIOS as version 1002 of Motherboard P4S8X-X, the specifications for your HP box says it's something else - a P5LP-LE.

In either case, your BIOS version is out of date.

Lastly, the answer to your question lies in the "staging" of the L2 Cache. - Or you could say that the access granularity results in the "increased latency". Technically, the test is detecting the cache correctly, and incorrectly depending on how you look at it. Obviously in your CPU there is more than one stage of the level 2 cache, or there is an access granularity issue, where only a certain amount of L2 cache can be pulled from one access method.