E2140 or E2160

[WEW]

If there is a like a $3 diff between the E2160 and E2180 - is it worth it to get the E2180? thanks

 

[quadomatic]

Originally posted by: WEW
If there is a like a $3 diff between the E2160 and E2180 - is it worth it to get the E2180? thanks

It would probably be easier to oc the E2180 since it has a 10x multiplier (the E2160 has a 9x multiplier). It depends on the motherboard you have; if your motherboard is good enough, you could push fsb further, and so it wouldn't really matter. Some people have said that E2160s OC the best out of the E2xxx series; whether there is any factual evidence to back this up, I don't know.

I would go for the E2180 if I were you, just because it would be probably be easier to oc with the 10x multiplier.

What motherboard would u be using it with? If you're using a motherboard you would have to use BSEL with, I would recommend the E2180, because then you wouldn't have to push fsb as much to get the speed you wanted, since pushing fsb may be a problem on cheaper motherboards.

You can ignore anything past this point, since it's not much help to you
_________________________________________________________________

I personally went with a pretty darn cheapo route, and I may sadly end up getting only what I paid for, but we'll see. I bought the E2200/ECS 945GCT-M/1333 cpu/mobo combo from Fry's for $78.99 ($85 with tax). I'm hoping to BSEL and do a Voltage mod to the processor to reach just under 3.0ghz. I'd say I'll have done pretty well for myself having only spent so little money.

But, I haven't assembled the system yet, or opened anything, so I may decide to sell the ECS motherboard and buy a Conroe1333 to do BSEL on that. The ECS motherboard doesn't have an option to change fsb, but the Conroe1333 does. I could do BSEL and then raise fsb from there to achieve a higher overclock. But, alas, there's still a chance I wouldn't be able to push my processor past 2.9ghz, so it may just be a waste of time.

I originally planned to buy a E2140 and a IP35-E, but the IP35-E has been OOS for so long. I managed to miss the deal with the rebate this month too

 

[Mondoman]

OTOH, some motherboards may work better with a "standard" FSB such as 1066MHz or 1333MHz. With a "standard" 1333MHz FSB, the e2160 runs at 3GHz, which shouldn't be a problem. The e2180 would run at 3.3GHz w/1333MHz FSB, which might be a problem.

 

[WEW]

I have Abit IP35-E

Honestly I would jump on the E2160 - but waited till Jan. 20 for the $64 price drop - which I have not seen reflected

 

[BonzaiDuck]

I'm PRIME95-ing my E2140 at 334 Mhz CPU host-frequency at 4:5 with RAM[DDR] @ 835 and 4,4,4,10,2T.

I set the vCore voltage to 1.3500V -- the BIOS monitor reported it idling at 1.345 on "Auto" before I fixed the setting.

How do they bin these things? For the price, I'm a bit stunned: from 1.6 Ghz to 2.6+Ghz is still amazing. How long before the light-bulb burns out on this sucker? If there's no "electro-migration" problem from the voltage setting, then there's temperature.

But I have the stock cooler [albeit with IC Diamond paste], and at room-ambient of 63F, the cores aren't breaking 46C. Now I thought -- even with the latest CoreTemp v.96.x, that the idle values were a bit low. I THOUGHT I saw them showing about 3C below room-ambient. Now suppose I use these settings during the summer, with room-ambient pushing 78F. Add 9C and add 3C for the error. The temperatures wouldn't go above 58C -- or 60C would be a stretch.

With the FREAKIN' STOCK COOLER?!! Are you KIDDING??!!! :laugh:

I can also see how to build a simple motherboard duct around the stock cooler, and I'm betting on another 5C temperature-abatement . . . . with the STOCK COOLER.

It either takes the fun out of spending money, or adds fun to spending less.

 

[BonzaiDuck]

. . . . So, then, with assurance that I'd done everything necessary in "cable-management," I replaced the computer case-shroud . . . .

The CPU temperatures went DOWN 2 to 3C, and "no ducting yet" . . . . .

That stock fan is just a tad noisy, though . . . . :laugh:

 

[WEW]

I did a price check today between the E2140 and E2160

Zipzoomfly has E2160 for $79 with shipping and the E2140 for $71.50 with shipping.


I have a Abit IP35-E motherboard - with the HP Ram that goes on sale -- plus Antec Earthwatts PS.


I'm looknig for best bang for the buck - I want to OC to 3ghz.

Am I better off saving $8-$9 going with E2140 or just go with E2160.

Thanks - lastly i want to use stock cooler - i have arctic paste though at home.

 

[BonzaiDuck]

Well, they're both 800 Mhz-FSB processors.

Somehow, I'm thinking that Intel is "dumping" these on the market, like Honda "dumped" the Civic-1200 1977-thru-1979 on the American auto market-- causing Lee Iacocca to go on TV and point fingers at the Japanese. What could AMD say in this market, when there are only two basic competitors as opposed to ten or twenty automakers world-wide?

They've managed to achieve a "level-of-quality" in the manufacturing process, they've probably reduced manufacturing costs at the same time, and they've pulled resources away from testing and "binning" the chips.

Thus, people on this thread suggest that it's a crap-shoot about how far you can OC these things. If the binning process loosely distinguished between chips which Intel decided had 0% chance of failure at the respective rated speeds, you might not require a lot of luck to get an E2160 to punch up to 3 Ghz. Then again, maybe some luck required.

For me, getting an E2140 up to 2.66 Ghz at the stock maximum idle-voltage setting is just like gourmet-pie-a-la-mode with the best French-Vanilla ice-cream and a bing cherry on top. I'm happier'n-a-pig-in-poop.

How far would you have to go to get an E2180 to 3 Ghz? You'd need to achieve a 50% over-clock. With my E2140 @ 8 x 334= 2.67+ Ghz, I've made a 60% over-clock without breaking a sweat. The Vcore is set to 1.3500V, showing closer to 1.36V idle in BIOS monitor. To me, that doesn't count for being "outside the voltage spec," and the motherboard "auto" setting gives me about 1.346V at idle. I still need to experiment with dropping it down a notch to see if it fails: it's gone nearly 6 hours as I speak using the 1.35V setting.

So I'm speculating. I'll have a chance to TEST an E2180 after Thursday, when it and an identical ($65 . . . Ha!) motherboard arrive in the mail.

So for an E2160 or an E2180, I could say it's possible, and I'm optimistic. But to confirm it, look for posters somewhere in threads on this forum -- maybe there are enough to build a frequency distribution of 3+Ghz over-clocks for any specific chip. Otherwise, we're speculating with a dose of logic thrown in here.

Compared to my "Cadillac" Striker Extreme motherboard, the bad press it's been given, the limitations (not too bad, though), and the price -- this seems to be the real satisfaction for over-clocking: A $65 motherboard, and an $80 processor. With that, DDR2 memory is really cheap. Check the 2GB Crucial Ballistix DDR2-667 memory kit at NewEgg with $30 rebate. I think -- with a mobo spec'd at 1,333FSB maximum, and the processor that is beyond its FSB spec but within its voltage spec -- that might be a good match -- just not as flexible as DDR2-800 or higher for different CPU : RAM ratios.

And I still think -- for my combination -- DDR2-1000 is over-kill and a waste of high-end memory. So -- for a $30 rebate, two rebates per household per product -- I placed my order with no regrets.

Another thing. For the prices and other considerations, I don't believe in OC'ing to the extreme limit -- even if you had phase-change or just water-cooling. To me, it's about "balance." And this particular machine -- it's gonna be Momma's. Momma . . . . imagines things in her head to worry about, which most people would dismiss as riskless and absurd frets. Give Momma a day, and there will be about 25 worries and hours of wasted time. So I must make sure this thing doesn't fail too soon.

[I think they might call it "Acute Anxiety Disorder," but people cling more to denial the older they get. ]

 

[lopri]

If you don't mind limited upgrade paths, I'd think DFI 975X board would be a perfect match for these chips. Performance :thumbsup: Features :thumbsup: Price :thumbsup: (usually $80~$90) That board isn't a value board so isn't built like one.

 

[BonzaiDuck]

Ya know? I'll put that in a "sticky-note."

A person gets too familiar with boards from a certain manufacturer, and real-world budgets, time and inclination make it impossible to get any other info except through hearsay. But "here-say" is usually reliable "hear-say," and I'll take a look at the customer reviews.

The 975 chipset was a good bet during 2005, and I gotta believe that the longest vision of an upgrade path for a board's chipset could only come from the manufacturer of the processor itself -- no matter how hard companies like nVidia would try . . .

 

[brencat]

Originally posted by: WEW
I have a Abit IP35-E motherboard - with the HP Ram that goes on sale -- plus Antec Earthwatts PS.
I'm looknig for best bang for the buck - I want to OC to 3ghz.
Am I better off saving $8-$9 going with E2140 or just go with E2160.
Thanks - lastly i want to use stock cooler - i have arctic paste though at home.

E2160 for sure, given its 9x multiplier. I cannot recommend you take a chance with the 8x multiplier of the E2140. While your OC goal of 3ghz is modest, there is always the possibility you will get a dud chip that walls ~ 340 FSB. If that happens, you'll be stuck at 8 x 340 = 2.7ghz. However, with the E2160, that same FSB would give you almost 3.1ghz.

FYI, the E2180 you see in my sig was only purchased b/c I got it for the same price as an E2160 at the time. With the benefit of hindsight, I think the E2160 M0 stepping would have been better, given that chip's consistently high OCs on much less vcore than the E2180. My chip is very average, and is a voltage pig above 3.2ghz. Thankfully, my Tuniq keeps it cool (~ 24c idle, 58c Orthos load)

You should be able to hit 2.8 - 2.9ghz on stock volts with the E2160/2180. Once additional vcore is required, it scales almost linearly @ 90 - 100mhz per 0.04 added vcore on the IP35-E. Additionally, Lopri has had great success OCing his E2160 to 3.6ghz (9 x 400) on a 680i mobo, and he's not the only one either. The E2XXX chips seem to like 680i for some reason.

 

[lopri]

Yeah.. I know we're talking about a budget build, but.. making choices based on $3/$8 is kinda.. well I don't know what to say.

In my case, I had the goal of 100% OC so I deliberately went with an E2140 first. It was dirt chip, too. (I just checked MicroCenter web site and they still list a retail E2140 for $60 and E2160 for $80.) But if you're looking for max OC then bite the bullet and spend that extra $3 for a higher multi.

 

[WEW]

Thanks guys - I finally just got the E2160 for $78.99 shipped from Newegg - I don't think its worth waiting for a few $ drop

 

[desertjedi]

Ewiz has a retail E2180 for ~$81 and an OEM one for ~$72. Mwave has a retail E2180 for ~$79.

I have an E2180 in my Abit IP35-E and it's capable of running 3.5Ghz on air (Zalman 9500). I currently run it at 9x360=3.24Ghz as I'm currently GPU-bound anyway. I'll see if that changes after I install my new 8800GTS 512.

 

[BonzaiDuck]

It seems I overlooked that the E2160 and E2180 have higher multiplier settings than the low-end E2140.

This is a very important factor in OC'ing, except that there is a manufacturing process and a "binning" process. Those chips may all come off the same assembly-line until they reach some point, where some become "8x" and others become "9x." Here, I'm speculating, but I am embarassed that I completely forgot about these differences in my big tome posted earlier, and it is important in one way or the other.

There was a time during end of the last decade when Intel was producing Pentium II processors. They had decided that they'd all come from the same manufacturing process, and that they'd all be designed to work at the speed of the "flagship" top-of-the-line model. The design incorporated a way in which the processors were "disabled" to run at slower speeds at the end of the "assembly-line," and then binned accordingly as "266 Mhz, 300 Mhz, 350 Mhz and 400 Mhz."

This enabled counterfeiters in the Phillipines to acquire the lower-priced 266 Mhz P-2's, and "re-enable" them so that they could be sold as "400 Mhz" processors. And at one time, Federal Trade Commission was eyeballing Intel and they had someone monitoring the forums at the Intel web-site.

That sort of thing is called "price discrimination," not too different than a doctor who charges a rich patient $5000 for an appendectomy and a poor patient $1000. But in the manufacturing case, people willing to pay less got the same product disabled to provide less, and those willing to pay more got the same product enabled to provide more. The economic analytic geometry shows that this sort of practice generates excess profit just like a monopoly on a single, indistinguishable product. But whether or not "product discrimination" is really something that FTC or DOJ Anti-Trust can go after -- that's a foggy, murky area.

 

[brencat]

Originally posted by: BonzaiDuck
This is a very important factor in OC'ing, except that there is a manufacturing process and a "binning" process. Those chips may all come off the same assembly-line until they reach some point, where some become "8x" and others become "9x."

I'm not sure there is much binning, if any, going on with these E2XXX as there seems to be a fairly wide distribution of OCs and vcore required across each of the main 3 (E2140/60/80). From my own observation of many posts & sigs in AT and elsewhere, the average OC seems to be 3.1 - 3.2ghz +/- 200mhz. Max FSB for most seems to be in 375 - 400mhz range. A few good ones can manage up to 425ish. But the one thing that has varied widely among these chips appears to be the vcore required for OCs above 3.1ghz. I know mobo & bios revision have a lot to do with that also, but have a look at the E2XXX/E4XXX OC thread by Zach where the vast majority of us have the Abit IP35-E and you'll see what I mean.

 

[BonzaiDuck]

First -- correction to a previous post I made: " . . . . But whether or not 'product discrimination . . . . " I meant "product differentiation." I'm getting old, and the jargon I learned in my earlier, professional life seems to be fading: "Dave?" "Yes, Hal." "Dave? . .. . my mind is going . . . . my mind is going . . . . I can feel it . . . . I can feel it . . . " [2001, a Space Odyssey; Arthur C. Clark/Stanley Kubrick]

Now, for Brencat:

We're not inside the Intel factory, so we're speculating somewhat. [or a lot.] They have to grab some processors and make them "9x" or "8x" and label them "E2180" or "E2140." We just don't know how they perform those "product flavorings," except that if they warrant an item for 3 years, they've had some sort of test-lab effort to prove that the chances of its failing at spec are essentially 0%. They would refine the process so that the process itself -- without much testing -- ensures a "no fail" rate.

Now, you may be "in the ball-park," and it depends again on how the chips are manufactured. They may have defined to process to virtually eliminate all chances of any unfinished product-unit failing at the E2180 spec within the 3-year-warranty time-horizon. So maybe they just roll those things off an assembly line, and tweak them in one way or another to make them "2180" or "2160" or "2140." And that information has to be "flashed" or encoded in the processor -- we know that.

 

[brencat]

Originally posted by: BonzaiDuck
[2001, a Space Odyssey; Arthur C. Clark/Stanley Kubrick]

I liked 2010 much better.

"...All these E2XXXs are yours...Use them together. Use them in peace..."

 

[BonzaiDuck]

Oh. The one with Roy Scheider. LOL. I thought at first (still waking up and no coffee) that there was some "E2010" processor I didn't know about! LOL

 

[Tullphan]

I had good luck w/my 2 gigs of $5AR HP ram.
I'm using an IP35-e motherboard & an E6300 cpu.
I bumped it up to 400mhz w/a 1:1 ratio & left the stock voltage alone.
Ran Prime95 & Memtest w/no errors.
I think I may try & tighten up the memory timings since they had no errors @ 1.8v @ 800mhz.

 

[BonzaiDuck]

What was the SPD stock timing spec-set for DDR = 800 Mhz?

I might be able to give you some tips if you post them.

The E6300 was always a very OC'able processor. Usually, it's the low-end that gets the rave notices.

I just slapped together an E2180, GA-73VM-S2 v.1.0 Gigabyte mobo and some 2x1GB Crucial Ball'x DDR-667 (published timing spec: 3,3,3,12).

I tried, as with the E2140, to punch it up to host-frequency 334 Mhz and 1:1 with memory, but it wouldn't post at the INtel retail maximum VCORE setting 1.3500V. Dropped the multiplier to 9, it looks like a slam-dunk for 3.0 Ghz.

My guess is that the VCORE has to go up some to make 3.33 Ghz.

The stock multi for E2180 is 10; for the E2140 it's 8. I'm deducing that it's 9 for the 2160.

Can't tell if the 2160 might be a better bet. Oh. I mean the E2160 -- not the year when our oil runs out and both the poles are melted . . . . . :laugh:

 

[BonzaiDuck]

O--KAAYYYY.

Here's what I think is going on in Intel's manufacturing process, and I'm sure that a lot of people had assumed this all along.

They produce a single CPU -- call it the "E21x0." Then they may test them all to make sure they run at 1.6 Ghz, and "do stuff" to the processors, and they go into a big box labeled "E2140."

Of those, they test a bunch to see if they run at 1.8 Ghz. They "do stuff" to those, and they go to box labeled "E2160." Those that fail go back to the "E2140" bin.

They take a bunch of processors from the E2160 box and test to see if they run at 2.0 Ghz, and the "do stuff" to those -- except the ones that fail, and they go back to the "E2160" bin.

All this time, the bidnis and cost-accounting folks are looking at the size of anticipated sub-markets for the E2140, E2160 and E2180, so they attempt to determine in advance how many E21x0's they need in each bin per unit time, or for each shipping cycle -- in order to maximize "profit." You know -- phase 3 in South Park's episode about the underpants gnomes? Phase I = "steal your underpants," Phase II = "???" and Phase III = "Profit???"

Here, I think we're speculating about Phase II . . . . "???".

Basically, ALL these processors are probably designed, and the process of manufacturing is fine-tuned -- to assure that they all run at 2.0 Ghz -- for whatever multiplier -- 8,9 or 10. Perhaps fixing the multiplier is something that's done toward end of assembly/fabrication.

That's maybe why I have to twist up the VCORE to get to 3 Ghz with an E2180 @ multipler = 9, but I can always get it to 2.67 Ghz at the same voltage setting I use for the E2140. And ALL of these boxes are labeled "Maximum Voltage = 1.35V".

Tell me I've got it wrong. I want to know more. Enlighten my silly a**. . . . .

 

[BonzaiDuck]

Oh, yeah. Skip the notion of "100% inspection." They ain't gonna do no 100% inspection for no $80 processor.

They have "production batches," and they random-sample each batch. Maybe they look for a +/- 2% confidence-interval on failure and choose their sample-size accordingly, estimate how much of a batch exceeds their acceptable failure rate ( the bidnis and accounting people will want to say something about "warranty RMA returns" or OEM-shipment returns), and if the batch no good, it goes back to the E2160 or E2140 part of the assembly line, multiplier is locked there, and they're packaged for shipment.

NOW I WANT TO ASK INTEL FOR A FACTORY-TOUR. So what're they gonna show me? Will it be like tourists to old Soviet Moscow, where you only get to see the "bright side?"

 

[BonzaiDuck]

Mmmm. Um-HMMMM. Interestin'. Verrry interestin'.

9 x 334 = 3.0 Ghz; DDR = 668 @ 1:1, latencies as spec'd @ 3,3,3,12 -- "no cigar" with voltage at 1.348V -- MEMTEST86+ errors. 10 x 300 = 3.0 Ghz, DDR= 600 @ 1:1 latencies of 3,3,3,12 -- so far . . . . so good.

Next it will be interesting -- innerestin' -- to see if the E2140 can punch up this high at 9 x 334.

Right now, preliminary voltage settings are just 0.001V above what BIOS reports with the "Auto" setting at stock 800 FSB . . . . set and "fixed" at 1.3500V. Both processors, E2140 and E2180.

I'd say it's an opportunity -- to get a few hundred clams from the fam-dam-ily as reimbursement, so I can play with both ends of the E21x0 processor line. Fun'n'games . . . fun'n'games. . . . . .

 

[lopri]

I don't think they can afford to do samplings. All CPUs must be tortured and pass through many hours of tests. Binning is a different procedures, though. Normally what Intel would do is grade a level lower than what they internally rate as a safety measure - in other words, if the chip is tested good to be an E6700, it'll be labeled E6600. If it's good to be an E6800, then label it as E6700, etc.

But with Core 2 Duo having so much headroom (thanks to AMD's incompetence), I guess they could simplify binning for bargain-bin chips like E2xx0s. I'd venture to guess that they validate all these E2xx0s @2.4~2.6GHz, then label them as per market needs. If Dell needs more E2160s, they just stamp more E2160s and if HP needs more E2200s then stamp more E2200s. If NewEgg sells bunch of E2140s then maybe they can label those as E2140 and order more retail boxes..