Thanx, Chilean.

Perhaps my question is better re-phrased as:

How much (performance) difference is there between a dual core cpu which has "Net Burst microarchitecture" HT, versus a dual core cpu which has "Nehalem microarchitecture" HT ?

Well, first of all, the "Net Burst" architecture wasn't really successful for Intel at all... remember how much better the AMD Athlons 64 were vs the Intel Pentium 4s? Well, that was because Intel chose the Net Burst architecture. To resume it down, the Net Burst architecture was based on the ideology that "THE HIGHER THE CPU FREQUENCY = BETTER PERFORMANCE" while this is true having all other variables constant, it isn't so if the CPU does less instructions per cycle. This is why, if you compare a 3.4GHz Pentium 4 with a Core 2 Duo running at 2GHz, the core 2 duo will beat it right out of the water (@ single core tasks).

Now regarding the HT, all Hyper Threading does is squeeze two threads into a single core. Which will give you a performance gain of LESS than 100%. Having two physical cores each running one thread is far better than having one running two threads. In other words, all HT does is multiply the number of PHYSICAL cores by 2. So if you have 8 cores (such as the most expensive i7), you will have 16 threads (the OS will see 16 cores, so it'll assign 16 different threads, one to each "core". In the case of Rosetta, 16 WUs will be running at the same time!)

You asked the performance gain, well by only knowing one has net burst and the other nehalem, that already tells you that the nehalem is better (it's far newer). But I think you are asking about the HT technology alone, I don't know in detail, but I'm pretty sure Intel reworked it's HT technology before putting it back into it's newer CPU architecture.

---

Thanx, Chilean.

Perhaps my question is better re-phrased as:

How much (performance) difference is there between a dual core cpu which has "Net Burst microarchitecture" HT, versus a dual core cpu which has "Nehalem microarchitecture" HT ?

Easy answer to the question of whether to HT or not, YES DO IT! I don't know of anyone that did it and is putting out less workunits as a result. BUT the percentage increase is dependent on alot of different things, including which projects you crunch for, so the amount of increase is variable from 25% to 50% or perhaps a little more. You will not see a 100% increase if you turn HT on, too many shared components that Boinc seems to use, slowing the crunching down as the cpu loads and unloads that stuff from one unit to the next.

Now if you are looking to buy a new pc, MaximumPC has a new 'dream machine' that uses 2 Xeon chips in it to give you a 24 core pc when HT is turned on! Now it is $15,000.00 too, but of course comes with ALL the bells and whistles, like 3 brand new Nvidia model 480 gpu's all hooked together and liquid cooled! The motherboard itself comes with [b]7[b] pci-e slots!!!

Already own two of those type of motherboards:

EVGA X58 Classified Four-Way


And one with 6 pci-e slots:

Asus P6T6 WS Revolution


Picked up (for a song, MUCH less than 15k!) a Dell 6850, 4P Xeon 7020, 32gb. Never heard of memory mirroring / memory raid b4 purchasing this thing!!!

The motherboard itself comes with 7 pci-e slots!!!

---

Defeat Censorship! Wikileaks needs OUR help! Learn how you can help (d/l 'insurance' file), by clicking here. "Whoever would overthrow the liberty of a nation must begin by subduing the freeness of speech" B. Franklin

Already own two of those type of motherboards:

EVGA X58 Classified Four-Way


And one with 6 pci-e slots:

Asus P6T6 WS Revolution


Picked up (for a song, MUCH less than 15k!) a Dell 6850, 4P Xeon 7020, 32gb. Never heard of memory mirroring / memory raid b4 purchasing this thing!!!

The motherboard itself comes with 7 pci-e slots!!!

Good lord! What do you do with that thing besides crunch?

Thanx, Chilean.

Perhaps my question is better re-phrased as:

How much (performance) difference is there between a dual core cpu which has "Net Burst microarchitecture" HT, versus a dual core cpu which has "Nehalem microarchitecture" HT ?

I've done some tests on my systems about 3-4 months ago, I've measured granted points differences/same cpu time between an Athlon X2 4850e, Intel E5200 and an old Pentium D Smithfield all clocked at 2.8ghz
same memory for the three systems

here the results based on average of seven difference kind of WU

Athlon has a performance ratio of 1.2/1.3X over the Pentium D, the Intel E5200 instead does 1.7/1.8X

from these many tests I've seen that HyperThreading helps the Netburst architecture a little bit, without HT those multipliers rise to 1.5/1.6X and 2/2.1X

Considering difference in IPC between Netburst and Nehalem with HT ON, the multiplier factor would be roughly 2.2/2.3X