wrote:Cedar Mill can achieve 5Ghz though.
Maybe on LN2 🤣 but it would probably still get smoked by a Pentium M on LN2 at half the clock speed. 🤣
who does that tho? 😵
OK guys maybe I am misremembering. Maybe some day I'll have the time to put it to the test personally.
5 GHz Cedar Mill is not that hard... with some luck you can get that clock on air or water cooling. But even then it is slow, especially in games. For example in my tests stock C2D E6300 (1.86 GHz) is 35% faster than 4.8 GHz P4 651.
HW museum.cz - my collection of PC hardware
higher clocks doesn't alleviate engineering malpractice of designing a 31 stage pipeline, compounded with bad branch prediction creating a marketecture to sell CPUs to consumers based on clock speed alone while ignoring poor thermal design and cpu efficiency. Shit CPU design. 🤣
For example in my tests stock C2D E6300 (1.86 GHz) is 35% faster than 4.8 GHz P4 651.
I think it's not that bad. And, In all fairness, it needs to be compared to 2M cache Conroe.
I must be some kind of standard: the anonymous gangbanger of the 21st century.
^GPU used : GTX 780 Ti
OS : Windows 7 x64 SP1.
4,9GHz on P4 651 required BIG air cooling but was game stable.
And part two of above table :
It's optimised for multicore support benchmark though. GPU driver is also optimised.
I must be some kind of standard: the anonymous gangbanger of the 21st century.
Multicore support and "driver optimisation" ?
I'd say it's irrelevant, since both Pentium 4 and Core 2 Duo's have exactly two threads, and they support almost the same instructions (Cedar Mill simply lacks SSSE3 while Core 2 - Hyper Threading).
SMP is not the same thing.
I must be some kind of standard: the anonymous gangbanger of the 21st century.
So ?
3DMark 11 should be optimised for both SMT and/or SMP (utilising whichever is available).
I just follow the rules : A Pentium 4 Cedar Mill and Core 2 test.
PS. You can see how 4 threaded OC'ed Pentium D's (ie. CPUs with SMP + SMT), are coping with two threaded Core 2s (SMP only).
EDIT : Older software is nice, but that would only make it harder for Pentium 4 to win (since Core 2 does a much better job at single threaded workloads).
Pentium Pro>P2>P3>tualatin>Pentium M>Conro>core micro architecture
Pentium 4 is not architecturally related to the above, its its own architecture.
were beating a dead horse thats been dead since 2003 with AMD murdered the Pentium 4 with the Atalon64, and then 2004 when Intel murdered the P4 with its own spiritual Successor to the tualatin.... the Pentium M.
The fact is P4 was only top performance for about 2 years and even that is disputed by people who were running SMP setups at the time in the early days of P4s where Dual PIIIs were still trashing them, and fast tualatin builds were still wrecking P4 in single thread.
Then about 2002 Intel held the crown but 1 year later got rekt by AMD hard and it wasn't until Conro came out intel even had a leg to stand on.
