When conroe hits the market will the real benchmarks:

[gman01]

When conroe hits the market will the real benchmarks:
-Perform 'Less' than the pre-release hype
-Perform 'The Same' as the pre-release hype
-Perform 'Better' than the pre release-hype

 

[gman01]

After 10 votes we seem to have a pretty good balance of opinions....

 

[Ycon]

Voters:
- Perform "Less" than pre-release hype: DVD, YO, LMM/BM/9nm/MrsD/BS/RB (heck, I dont know how many accounts that guy has)

:lol:

 

[gman01]

After 18 votes it seems 83% are conroe believers, and 16% are not....

I have to give 1 point to the intel PR department....

 

[the_vorlon]

It is not a matter of believeing, performance is a matter of reality...

There have been a great many reviews already done where actual Woodcrest chips have been benchmarked, and they perform they way they perform...

There are always modest tweeks to the die as subsequent steppings come out, but these are tweeks - NOT anything radical.

It's not like Intel produces some fake Woodcrests that go like $%^%% and the will produce bogus ones that don't..

Hey the shipping ship IS the engineering sample +/- a tweek or two...

Performance won't change.

Intel fanboys will find and tweek specific benchmarks that show Woodcrest to be a killer monster chip (SSE optimized, fits in 4 meg cache)

AMD fanboys will find apps where Woodcrest hurts (Huge bandwidth, non-optimized code with lots of code branches) to show it sucks...

But broad specturum benches will show what they have always shown...

 

[MU_Engineer]

I'll bet that the benchmarks that Intel let people run on its ES black-box systems will perform pretty much the same on the real things. It would be marketing suicide for them not to. However, I bet that the OVERALL performance of the chip will not be as great as those cherry-picked benchmarks would lead you to belive. A manufacturer's not going to let people run a benchmark that does something the CPU is not that good at if they can help it, you know. They're going to have the thing that the CPU does best be on display to wow people.

You just have to think a little here and you can figure out what will happen.

 

[quantumsheep]

I believe that they'll either perform the same, or better than, the pre release benchmarks. The reasoning behind this is that Intel will have improved the production method etc.

 

[jaguarskx]

Since there have been several independant hardware review sites that that previewed Conroe recently, I would say that Conroe more or less meets the hype.

 

[RichPLS]

I think retail Conroes will perform similar to the latest reviews (past week or two) which are using the latest chip revisions... as they actually show some performance gain from the older revision reviews...
Other than that, I feel it is clear that Conroe performs as expected...

 

[lbax]

I have seen enought benchmarks, from a wide range of sources, to lead me to believe that Conroe performance will be within the published envelope of expectations.

 

[gman01]

I hope the majority is correct, and conroe lives up to expectations.... We need a cpu war from these two companies.... At the very lease AMD needs to fix/get their ddr2 memory controller running optimumly.... Also, AMD needs to start cranking out quality quad core starting yesterday....

 

[49ers540]

The reason behind pre-release benchmarks, is to build up hype, and Intel has done that. I think benchmarks will be about the same, maybe little lower.

Woodcrest, shows better benchmarks. :lol:

 

[MU_Engineer]

Not in Apache or scientific apps according to GamePC.com, who tested a Woodcrest 2.0 and 2.33 vs. a Dempsey 3.73 and Opterons 1.8-2.6 GHz. The Woodcrest does well in encoding and rendering, but it is not nearly as good of a webserver as an Opteron rig. Encoding isn't really something that many do on "big-iron" machines, and there are some render farms, but scientific computing and especially web serving are what servers are for. I have not seen terminal server/thin-client-host or database benchmarks, and those things are some other real uses of servers.

 

[geralt]

Good post.
Hopefully after monts of waiting we will finally see 64-bit performance.

 

[quantumsheep]

Ahh but companies that use huge servers dont ALWAYS look at benchmarks, they will look at the power usage. As using lots of CPUs at the same time will use a HUGE chunk of power (My D805 O/C rig consumes a huge chunk of power, imagine running 100 of those in sync?) Any saving is a major plus on any CPU.