Intel's Pentium 4 670: Just Another Speed Bump
by Derek Wilson on May 26, 2005 9:00 AM EST- Posted in
- CPUs
Final Words
So, what is the final verdict? There are cases where the 670 performs better than the 570 and cases where it performs worse. It all comes back to the differences between the two processors. Running with the Pentium 4 670, some applications benefit from having more cache, and other suffer due to the higher latency.To recap, applications that stream data (encoding and rendering) aren't impacted by the cache differences between the 5xx and 6xx series. The two processors perform nearly identically in the tests that we ran in these areas. Games have generally seen a boost from the added cache (we've seen this since the first Extreme Edition), so the 670 has a slight edge over the 570 in that category. Under office and workstation applications, the impact varies from case to case.
The Pentium 4 570 comes in at around $645. In quantities of 1000, the 670 is $850, and the only one listed on pricewatch right now is over $1000. If high single threaded performance is required and the choice comes down to the 570 or 670, we'll have to recommend the cheapest one. Right now, that's the 570.
We do have trouble recommending either of these processors for their price, especially when less expensive options don't degrade performance.
We didn't talk much about the Pentium D 820 today, as Anand covered its performance characteristics in his earlier article on the subject. For those interested in the performance of that part, please check it out. As far as recommendations go, the Pentium D 820 is a compelling solution for those who don't need huge single threaded performance. The choice between great single threaded performance and solid multithreaded performance with Intel's dual core processors will really come down to the usage scenarios that fit each individual.
All in all, we can't say that the Pentium 4 670 adds anything to Intel's line up. As higher performance dual core processors start to come along, we will certainly be more interested in them.
Facebook
Twitter
RSS
33 Comments
View All Comments
bob661 - Friday, May 27, 2005 - link
Since you're here Jarred. Was the tested 4000 a San Diego core? Thanks much.bob661 - Friday, May 27, 2005 - link
#16There ARE San Diego core 4000's. Check here:
http://tinyurl.com/cdy8m
JarredWalton - Friday, May 27, 2005 - link
Certain benchmarks are not 100% repeatable. WinStones, SysMark, WorldBench, etc. can all vary by a decent amount. While running multiple benches does help a bit, you can still end up with some odd results. I've seen variance of 5% on some benchmarks, for example. I don't know about the WinZip and Nero results, though - it looks like some other hardware or driver provided for a major change.flatblastard - Thursday, May 26, 2005 - link
#10 "don't forget, a 200 MHz increase with AMD cpus is like a 300+ MHz increase for intel"#13 Also consider the fact that 200 "A64 Mhz" aren't equal to 200 "P4 MHz"
I am aware of this, and I will now make you aware of the fact that I can afford to have 10 less FPS.
Tujan - Thursday, May 26, 2005 - link
""Or does an endorsement by the Blue Man Group wash away all sins? :) ""............Nay think its a matter of reviewers not being aware of when the next shipment of bananas is going to come in.
A load of bull can be a load of bull sometimes.
Tujan - Thursday, May 26, 2005 - link
Is lower or higher 'better for the Sysmark Data Analysis scores ? The 670 got 183"". [ ]Weird how the AMD Athlon 3400+(2.4/512/1ch) did so badly in the Specviewperf 8- Pro/Engineer Performance Engineering and SolidWorks Viewset....
I was trying to tell how well a given processor would do,so bliping on the AMD Athlon 3400+(2.4/512/1ch) I thought that would be a good processor to have,now is that a 754 platform processor/motherboard.
Couldn't find same range processor from a retailer in 939 wich would only show it in 754?
[ ]
Somebody here said 1ch/2ch is dual channel ? The benchmark setup did tell of using the 754,lest an Nvidia SLI is one of those ?
SLIM - Thursday, May 26, 2005 - link
Heres another request for some kind of explanation about the very odd scaling amongst the 6xx series chips especially in the PC Worldbench results.mjz - Thursday, May 26, 2005 - link
14 - the 4000 is an FX 53 without the multi.. it is not a san diego core.bob661 - Thursday, May 26, 2005 - link
Also, could you guys include an Autodesk Inventor or Mechanical Desktop bench?bob661 - Thursday, May 26, 2005 - link
Does anyone know if the 4000 used was a San Diego core? Thanks.