CPU info overload, help please.

[wolfwatcher51]

Have been reading the various topics on this forum and got a lot of info. First computer build. Thanks to the members, I have selected the ABIT IP35 MB, Antec EA500 PS, and GSkill F26400CL5D-4GBPQ ram.

Seems reasonable to get 2x2GB rather than 4x1GB so I have two slots open for future X64 upgrade to 8GB. OK? I think I saw where two sticks is easier on the memory controller than 4 sticks.

Anyway, to the problem. A link from here to Tom's Hdw got me to an article from 9-12-07, "$ 89.00 Pentium Dual Core that runs at 3.2Ghz". They used a lowly E2160, 1.8Mhz,1MB L2Cache,9x,and 800MHz FSB CPU and cranked it up. You could use a $22.00 ACFreezer7 Pro, and it would be chiilly.

hey then tested it against the E6750 and the difference was minimal. except the E6750 cost more than 2X and requires about $60.00 worth of HSF. I believe the E6750 was run stock, but not sure now. Yes, I realize the E6750 can go way up there also, but wow. I just do not understand how the E2160/1.8GHz/1MB L2Cache/800FSB/9x multiplier could test so close to the E6750/2.66MHz/4MBL2Cache/1333FSB/10xmultiplierwith G0stepping!

What is going on here? One thing is they did have to raise the FSB above 1333 to 1420, and the Vcore was 1.525 volts to get the 3.2GHz. Is this OK to do to the IP35? Is 1.525volts a life killer to the chip?

Even if they left it at FSB 1333 they still got 3.0GHz. For $90+$22 bucks? Please say it ain't so, show me the trick Mr. Wizard, please. This has me considering the E2180 for $89.99 at NewEgg rather than the E6750 for $189.99 for now, and then get one of the new chips this time next year. Thanks for your help, in advance, Chris.

 

[lyssword]

Why it would test so close? because 3.2 is more than 2.66ghz thxbye

Jesus that's one mess of a text block.

 

[VirtualLarry]

I just put together a rig with a GA-P35-DS3R and an E2140 (1.6Ghz), and have overclocked it to 3.28Ghz on 1.425v vcore. I don't know why TH had such high vcore, I wouldn't recommend going as high as they did. Luckily, in most cases, it's not necessary as my overclock has shown.
The E2140 is $60 at Microcenter btw.

 

[taltamir]

I cant read it... My eyes just start swimming... please edit in some text breaks in there.

 

[Billb2]

Originally posted by: wolfwatcher51Please say it ain't so, show me the trick Mr. Wizard, please.

No trick... that's what ovelocking is all about.
Two cavests:
Results may vary.
Think what ya could do with a E6750!

..my eyes hurt.......

 

[wolfwatcher51]

Lyssword,
So you are saying because 3.2 is>2.66, the chip will test and perform better?

One chip has 2 FSB steps faster and 4x the L2 cache, and that does not change the results?

I thought that if two processors had the same cpu speed, the one with the larger L2 cache would be faster, kind of like more ram on the motherboard. And, that increasing FSB would also speed things along.

So, what is the criteria for using an E6750 at 3.2, vs an E2180 at 3.2, besides the extra oc still available on the E6750?

 

[Quiksilver]

I'm sure if you were to have an E2180 and an E6750 clocked at exactly the same speed with the only difference being the L2 Cache amount that the E6750 would win overall.

 

[myocardia]

Originally posted by: wolfwatcher51
hey then tested it against the E6750 and the difference was minimal. except the E6750 cost more than 2X and requires about $60.00 worth of HSF. I believe the E6750 was run stock, but not sure now. Yes, I realize the E6750 can go way up there also, but wow. I just do not understand how the E2160/1.8GHz/1MB L2Cache/800FSB/9x multiplier could test so close to the E6750/2.66MHz/4MBL2Cache/1333FSB/10xmultiplierwith G0stepping!

I'm assuming you're talking about this article? I see the E2160 getting beaten in every test by the E6750, even though the E2160 is running more than 500 Mhz faster.

 

[lyssword]

Originally posted by: wolfwatcher51
Lyssword,
So you are saying because 3.2 is>2.66, the chip will test and perform better?

One chip has 2 FSB steps faster and 4x the L2 cache, and that does not change the results?

I thought that if two processors had the same cpu speed, the one with the larger L2 cache would be faster, kind of like more ram on the motherboard. And, that increasing FSB would also speed things along.

So, what is the criteria for using an E6750 at 3.2, vs an E2180 at 3.2, besides the extra oc still available on the E6750?

Basically, fsb doesn't matter that much and also e2180 @ 3.2ghz will have around 1333fsb which would be same as non-oc'd 6750, because when you overlock you raise FSB as well.
L2 cache does matter, but only 5-20% at max. So it will be slower by 20 % lets say, but if it's 20% faster clocked, then it will match e6750 exactly or exceed where l2 cache is not as important

 

[Quiksilver]

^^^^ re-read the last 2 sentences in what you quoted.

He's asking if they were both OC'd to the same clockspeed, not if one remained stock and the other OC'd.

 

[lyssword]

Originally posted by: QuiksilverX1
^^^^ re-read the last 2 sentences in what you quoted.

He's asking if they were both OC'd to the same clockspeed, not if one remained stock and the other OC'd.

My bad, I typed that before he edited his text block. I got kinda lost in that jungle Actually he did say STOCK 6750 vs 2180 oc'd I re-read whole thing again. However, if both run at same speed (lets say 3.0ghz) 6750 could be 20% faster depending on app.
Also, to OP, e6750 would not run hotter than 2180 if both are at same speed. So you don't necessarily need $60 cooler.

I myself have e2140 runs on stock hsf stock volts @ 2.8ghz

 

[taltamir]

Well, those chips overclock amazingly well... I wonder how well an E6750 will overclock though... maybe you can take it to much greater highers.
As billb2 said, Getting lots more for your money is what overclocking is all about.

Keep in mind though that architecture improvements in new CPUs give many boosts to performance even in the same clock speed. SSE4 for example would double your encoding speed with supported applications (divx encoding for example).

 

[BonzaiDuck]

I just don't see what all the hoopla is about.

When the bottom-of-the-line Smithfield dual-core (810 . . or 805 ?) was tested by Tom's -- maybe in early 2006 -- they were touting its over-clockability and under-$125 price-tag.

Just about every bottom-of-the-line processor in each family over-clocked better than the top-of-the-line. Thus, the low-end Smithfield, the low-end E6300 -- or going back a few years, the low end 2.4C P4 Northwood. You could get the 2.4C to 3.0, but you might only get the 3.0 as far as 3.6. That's a 25% over-clock versus a 20% over-clock. Throw in the Prescott "E" processors, and it seemed they wouldn't go as far, but the L2 cache-sizes meant greater "scalability" at higher clock speeds, even if the higher clocks weren't all that stellar.

That speaks to Intel's production process and the nature of that business. None of us would be eager to pay the $1,000 price-tag for the flagship model so we could over-clock it, and in prior generation CPUs, the flagship model was not the most-overclockable in the line.

But with those benchmarks in the THG tests, I agree -- the E2180 "held its own," but didn't capture "first-place." It may be the L2 cache-size -- a primary difference reflected in the cost.

At my ripe old age, I fancy that we're like a bunch of 7-year-olds racing balsa-wood sailboats in regatta at the local public-park's pond. Dress up your computer-case, and it's no different than giving the sailboat a flamboyant enamel paint-job.

With the real-estate market going south, I also bet that the economy "needs" our obsessive over-spending on this stuff, even when we buy bottom-of-the-line or middle-of-the-line.

And those of us who shell out the $1,000 to $1,300 for a QX9650, are either taking a gamble that we can over-clock the thing into the next dimension to "where that ol' demon lives in the sky," or we just don't want to get our hands dirty with the time and trouble.