Is the 3.0 GHz Woodcrest worth it?

[bob5820]

An old rule of CPU purchase is never go for the top of the line. Ths best price / performance is always one or two clicks down. The 2.66 seems to be the wise choice here. Just one comment on upgrading CPU's, though it is often possible many time it just doesn't make sense. By the time I'm ready to upgrade my CPU (2 or 3 years down the line) its likely there will be other considerations such as higher FSB, different memory, new bus technology, etc.

 

[Scarlet Fever]

i read somewhere that if the speed difference is less than 10%, you won't notice any difference. The 3.0 is slightly more than 10%, but probably not enough to notice. As said before, the money saved would be better put towards more RAM for it. A 2.66GHz with 4GB RAM will fly past a 3.0GHz with 1GB RAM.

 

[killr_b]

Generally I'd say if you have to ask, then no. Top of the line anything is for the luxury.
2.66 w/ 4GB ram will for sure be faster in real world use than the 3 w/ 1GB.

I got the 3 because I could. If you have a certain amount of money allocate it in stages. Enough space, more ram, better graphics- oh look, money left over, better cpu.

I actually have my Mac Pro now. And I feel I got my money's worth. The price per performance was about what I was expecting because they always give the same incremental updates. The only real gip was the HD, but I needed a solid system on my desk by the 11th or I'd have to wait until I finisned that project to upgrade. So far absolutely no bugs.

Bottom line is, get the ram if you can't afford both.

 

[Glen Quagmire]

I doubt Clovertown debuts at a price that dethrones Woodcrest. Meaning Apple wont use them or offer them as a very high priced upgrade.

Clovertown won't be in a feasible price range til at least 2008.

At which point it will be replaced by the successor to the Conroe/Merom/Woodcrest architecture.

 

[miniConvert]

Absolutely not worth it. I thought long and hard about the choice myself and decided on the 2.66's. With 4 cores in the system and the prospect of applications becoming increasingly multithreaded there'll be no increase in the life of the product through opting for the 3Ghzs's, plus the higher clock speeds are likely to result in more power draw and more heat - possibly resulting in more noise. The speed advantages of the 3GHz chip will only be noticable to a vast minority of people who are either maxing out a single core with an application or who are performing extensive tasks loading up all four cores.

 

[Mundy]

All the mentions of Clovertown in this thread made me want to do a little research.

The consensus among a number of analysts, including one at Insight64, seems to be that Clovertown will not even be competitive with Intel's own Woodcrest architecture. Woodcrest has dual-independent front-side buses, while Clovertown will mark a return to a single, shared FSB. This is a similar architecture to Intel's first-generation Intel CPUs, which performed so poorly compared to the AMD X2 line that it forced Intel to make unprecedented price cuts.

It appears that while Clovertown will offer some benefits to highly multi-threaded apps of the sort used in finance (and other situations requiring constant database access), Clovertown is sounding less and less like a good desktop CPU. I can't find any indication that Intel developed Clovertown with workstations in mind; it seems to be squarely aimed at the dual-socket server market.

EDIT: I just read something which confirms this. Tigerton, not Clovertown, is the successor to Woodcrest. The original replacement for Woodcrest was supposed to be Whitefield, but that has been canceled. Moreover, Intel has not confirmed how many cores Tigerton will have (i.e. it could be less than four). It will be at least the second half of 2007 before Tigerton arrives. I don't think we will ever see Clovertown in a desktop Mac.

 

[Glen Quagmire]

Personally, I would go for a Woodcrest Mac Pro (as I have done!) and skip any future model that sports the successor to Woodcrest (whatever that may be). Woodcrest doesn't appear to suffer from any saturation/bandwidth issues caused by the FSB.

In 2008, Intel is due to release its next architecture - Nehalem - which succeeds Woodcrest/Conroe/Merom. Supposedly it will ditch the FSB and use a new interconnect named CSI. There might be an integrated memory controller as well, just like the 64-bit AMD chips.