• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

e6x50 vs. e4x00

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Originally posted by: zach0624
I know the 680i runs hot but this the mobo will need to last a couple of years without being replaced so it will probably need sli (don't like crossfire) and yes the 680i will support penryn.

If you're that concerned about performance, then you should never plan ahead, especially with a time frame of several years. Would you be happy today with an NF2/NF3 board humming along at 1.5GHz?

 
Originally posted by: SerpentRoyal
Originally posted by: JD50
Originally posted by: SerpentRoyal
If RAM speed is set to AUTO detect, then you're already running at 400MHz speed (DDR2 800) with 200MHz FSB! Use CPUz to check your RAM speed at the overclocked setting. It may show 600MHz RAM speed with 300MHz FSB.

E4300 can handle 1.45Vcore if you have a decent CPU cooler.

I'll have to check it when I get home, but I believe that CPU-Z is showing the RAM speed around 370mhz, which is really 740 mhz because its DDR2 right?

If I manually change it to a 1:1 divider, doesn't that mean that I would have to get my FSB up to 400mhz for my RAM to be at its stock speed?

Correct...your FSB = Memory speed with 1:1 divider.

So wouldn't it be bad to set the divider to 1:1 since the highest I will get my FSB up to is around 310 or 320? It seems like my RAM would be really underclocked at that point.

I think its automatically setting the divider to 5:6 right now.
 
You want to underclock your RAM to take it out of the equation. This is the easiest way to explore the maximum core speed of your CPU. Once you know the maximum stable core speed, use a different memory divider to increase the RAM speed. Run memtest to test stability of RAM.

Let's say your maximum stable FSB is 320MHz. If you use the 1:1.25 memory divider, then RAM speed will 400MHz (320 x 1.25). Quality DDR2 800 RAM rated at 1.8V should remain stable up to 440MHz with 2.0V.
 
Originally posted by: SerpentRoyal
You want to underclock your RAM to take it out of the equation. This is the easiest way to explore the maximum core speed of your CPU. Once you know the maximum stable core speed, use a different memory divider to increase the RAM speed. Run memtest to test stability of RAM.

Let's say your maximum stable FSB is 320MHz. If you use the 1:1.25 memory divider, then RAM speed will 400MHz (320 x 1.25). Quality DDR2 800 RAM rated at 1.8V should remain stable up to 440MHz with 2.0V.


Ahhh ok, I get it. Thanks for your help!
 
@serpentroyal the reason what I am trying to say is that I want a machine that can be moditeratley upgraded in a year to help extend it's usefull life.
 
SLI is a bum deal. To get the best performance, you'll need to buy two high-end GPUs. In six months, your SLI rig will be slower than a high-end GPU.

A P35 board has the best chance of overclocking the upcoming 45nm Intel CPUs. That's the most logical upgrade path in one year.
 
Originally posted by: SerpentRoyal
SLI is a bum deal. To get the best performance, you'll need to buy two high-end GPUs. In six months, your SLI rig will be slower than a high-end GPU.

A P35 board has the best chance of overclocking the upcoming 45nm Intel CPUs. That's the most logical upgrade path in one year.


I noticed that, it kinda makes SLI pointless doesn't it? I mean, by the time there are actually games out that would really challenge a high end GPU to the point that you would need SLI, the next generation is out. Unless you are doing it just to do it, it seems kind of pointless.
 
just get a E2140/2160 as a place holder. these chips easily oc to 3ghz+ with a small bump in voltage. thats what i'm planning on doing soon.

 
Originally posted by: SerpentRoyal
SLI is a bum deal. To get the best performance, you'll need to buy two high-end GPUs. In six months, your SLI rig will be slower than a high-end GPU.

Keep in mind that the 8800-series cards came out last year. So, it has been 9 months now that GTX SLI will have been going strong without a single card that can match it.

Granted, not too many games *need* that much power. And the ones that do, it's pretty much only at 2560x1600 (30" LCDs).

I don't know when nV's (or AMD's) new card is coming out. But my guess would be Nov-Dec timeframe. If so, and *if* the new card is twice as fast as the 8800GTX, then, after 12+ months, we will finally have a single card that can match GTX SLI. Still, the next gen card will likely have features that the 8800-series can't touch, even in SLI.

Granted, I'm not telling anyone to go out and buy 2 GTXs unless you really really need it (especially so late in the product cycle). My point is-- "six months" isn't an accurate estimate.

And those that did jump on GTX SLI back then will probably be able to sell both of their cards for $700 total, making their upgrade cost for next-gen SLI at around $400-$500 out-of-pocket (depending on what the new top end cards cost). So, while I'm plunking down $500 for a single next-gen card, they'll be plunking down the exact same and getting 2x the performance (depending on the game and settings).

I wouldn't call SLI a "bum deal". It has its merits. In my opinion, however, it is only useful when buying 2 top-end cards right away and playing at very high resolutions. Not something most of us can/will do.
 
Hmmm...not much advancement in technology because NVIDIA's closest competitor ATI is in the DOG HOUSE, thanks to the idiotic AMD acquisition. I have better things to do with my $ than soaking it into a pair of uber-expensive GPUs. Different stroke for different folks.
 
Originally posted by: Markfw900
Originally posted by: SerpentRoyal
Originally posted by: Markfw900
E6750.... WAY better than an E4xx0 series IMO.


Hmmm...you must be working for INTC's marketing. E6750 has twice onboard cache vs E4300. At the same core speed, 6750 is marginally faster with benchies...about 1 second with 1M Super Pi.

Real-world use? Perhaps a gain of a few FPS, but only with some games.

I can't justify paying twice the price for such mediocre boost in speed.

screw the cache. I just know the E4300's track record for OC'ing isn;t that great, and the worst E6750 is the 3.5 oe of bryan's, so I think the odds would be way better for an OC.
ok, ok, I"m going to go prove that I can get over 3.6 orthos stable. I won't run it like that all the time b/c I need over 1.5 v just to make it to windows, but I'll do it!

 
Remember the famous bell curve in high school? 50% of these chips with do 3.2GHz. 10% will hit 3.6GHz. And 10% with not break 2.8GHz. Is this so bad for a CPU originally rated for 1.8GHz? Most people will only bit ch if they can't get what they think is the norm!
 
ok, ok, I'm orthos stable at 3.6 at 1.54 v. My temps are 57c max with my tuniq tower running full blast. How high should I go before stopping the voltage madness? I would like to find the ultimate max oc, but I don't want to blow it up.
 
Under 65C as reported by coretemp 0.95. If the chip has 100C Tjunction, then you can push temp up to 80C for about 1/2 hour. Any higher and you'll probably encounter stability issue.
 
yeah, well, it does have 100c tjunction but it only shows 85c in coretemp. At least, I hope that it has 100c tjunction b/c otherwise it is violating the 2nd law of thermodynamics. hmmmmm, ambient 27c, cpu0/1 idle temps 19c at 3.68...

is there a max voltage that you would recommend for a short time oc, or do I just need to make sure that I keep it below 80c?
 
From your posts, I think you are already pushing the CPU well beyond the point of 24/7 stability. I'd say 3.5GHz is the max for this chip on air.
 
Back
Top