Slashdot Mirror
Intel Core 2 Duo E6750 Sample Preview
MojoKid writes "Intel took the wraps off a new Core 2 Duo desktop chip today, dubbed the E6750. Though this chip shares the same basic clock speed as the Core 2 Duo E6700 at 2.66GHz, this new processor also runs on a faster 1,333MHz Front Side Bus. The new chip's additional bus bandwidth affords it up to a 5% performance advantage over standard 1066MHz FSB-based Core 2 chips. However, what's perhaps more promising is this new chip's
overclocking head-room of up to 3.92GH and beyond on standard air cooling."
The link to the article all on one page is http://www.hothardware.com/printarticle.aspx?artic leid=989
Why not just sell them rated at a higher clock speed? It would be funny to think they made a fast chip and purposefully rated it at a lower speed to grab some of the extra hobbiest market while simultaneously cutting down on support calls from overclockers who cause system instability by making the overclockers think they are overclocking. :)
....a coincidence: the overclocking article is from ->hothardware. yeah. i truly believe, that an ordinary aircooling and a C2D at 3.92GHz have really earned an article on a domain called like that.
overclockability. I believe those chips sent out for review are cherry picked by intel. Most of the reviews for the Core 2 Duo chips last year said the lower end chips could easily be overclocked to 3.5ghz+. That ended up not being the norm. I think something around 3ghz is pretty standard.
I thought we had finally put the "megahertz myth" behind us. But no, here we are again cheering on Intel for producing chips with their many megahertzes and gigahertzes.
We should lean on them to use a more sensible naming convention. AMD has led the way in this area. Consumers are much better served with descriptive product names such as, for example, "Turion 64 X2 TL56", rather than some arbitrary clock speed designations.
This is like sex, except im having it!
Great, so it's bloody fast. But can it complete 10+ hours of Prime95 and 32M digits of SuperPI without any errors? Simply booting and running a few benchmarks is hardly a means of stability testing.
Now what I would like to see advertised - but won't - is slower but highly reliable motherboards, processors and memory at commercial prices. How about a Core Duo Reliability Edition? I would reallyt like to be able to build a server and a few desktops from commodity hardware and almost be able to forget about them for 5 years. I can get HDDs that will do that, but where can I get the commodity silicon where the manufacturer will make a statement about long term reliability?
Pining for the fjords
Do people still overclock? It is such a focus on this in online hardware reviews, but none of the people I know still do it, even the gamers. Power consumption, heat and noise is much more important to them. Low sample number to draw any significant conclusions from, I know, but still... Perhaps the market has moved on a bit?
Also, whenever they do speed comparisons, I wish they would add in models from one and two years ago. I really don't care if a chip is 0,05% faster than its similarly priced competition, I want to know if it is a good time to upgrade my old computer.
Being bitter is drinking poison and hoping someone else will die
Exactly, and why the chip is sold as 2.66GHz not the 3.92GHz that the marketing department would prefer. Semiconductor manufacturers do a stellar job of testing and specifying things over the complete operating range of the device. Ignoring obvious differences in things like ambient temperature and power supply fluctuations when you overclock a device you risk a number of factors for reliability. Any temperature measurement is always taken at a single point and if another point on the surface of the silicon is hotter, for example because your application of heatsink compound was not so great or it contains higher speed switching and more dense circuitry in that area you always run the risk of frying things. Not to mention there is a difference between running a game that might place peak demands on the CPU and allow it to cool versus compute-intensive applications where you might want to drive all cores at 100% over a long period. And they might be using a different section of the processor, and your CPU might be from a different batch, and...
Some models of Xeons run at 1066 and 1333. Just off the specs on the article I'd say they released those Xeon CPUs as desktop model on the LGA775 socket. The larger cache, higher bus speed, thermal design and Smart Cache match up to the Xeon E51xx and E53xx.
Of course people overclock. instead of buying the 2.6GHz Core 2 Duo, you just buy the 1.8GHz version and pay half the money ($160 insted of $320).
Now just overclock it back up to 2.6GHz.
You may want to do a little 2 corner testing (Voltage and Temp), just to make sure you are within stable regime.
As long as you dont overvoltage the chip, there is really no reason not to max out the clock rate. As soon as the CPU idles, it underclocks automatically anyway, so you get the boost only when you need it.
If you do any home video decoding, the difference is huge.
To make the point clear: You can burn out a power transistor if you run it too hard, but this is not possible on a CPU. It will hang long before it even gets close to be damaged. If the chip overheats and/or is driven at a too high clock, it just hangs. Reset and cool, and it is good as new.
don't cut it off www.mgmbill.org
When you are overclocking for 5% performance gains on synthetic benchmarks there is no time for commas.