i7-3820 vs. q9550 Sisoft benches: underwhelming

[magellan]

I had my q9550 running reliably @ 4.25Ghz (until it died when I tried to push it farther) and I have to take my i7-3820 to 4.8Ghz in order to double the Sisoft Sandra 2011 whetstone and dhrystone benchmarks (which takes 1.48V vcore and lots of heat), which is kinda underwhelming, especially considering the vcore required to get there. I used to think my q9550 LGA775 system was loud...until now. Even more surprising is that the Sisoft Sandra 2011 memory bandwidth benchmark only shows a 50% increase over my q9550 (while memtest shows a 300% increase). I used Sisoft Sandra 2011 because that was the only results I had for my overclocked q9550.

 

[GoD_tattoo]

So you joined in 2002 and only have 10 posts? That's in line for the lurker of the decade award...

On topic, I would say another benchmark might show a bigger improvement without such a high o/c. Can't really go off one benchmark. Who knows if that is updated to take advantage of newer processors? Not sure

I'm still amazed at the 10 posts in 11 years...

 

[EarthDog]

Wow, you are all over the map..LOL!

1. Loud is a function of the cooler strapped to the thing. If its too loud, get another.
2. Expectations were perhaps not set properly on the performance of the CPU
3. Sandra 2011 likely doesnt take in to account quad channel versus dual (not that it remotely matters anyway).
4. One benchmark does not a 'big picture' make!

It just seems like you had unrealistic expectations of the platform and its performance profile. Should have made what a lot would consider to be the proper jump and went 3930K on X79. 3820 is no different than a 2600k minus the ability to change the multiplier. Perhaps with a 3930K you would see your 2x increase on multithreaded applications? Not sure.

Sorry you are disappointed.

 

[magellan]

Wow, you are all over the map..LOL!

1. Loud is a function of the cooler strapped to the thing. If its too loud, get another.
2. Expectations were perhaps not set properly on the performance of the CPU
3. Sandra 2011 likely doesnt take in to account quad channel versus dual (not that it remotely matters anyway).
4. One benchmark does not a 'big picture' make!

It just seems like you had unrealistic expectations of the platform and its performance profile. Should have made what a lot would consider to be the proper jump and went 3930K on X79. 3820 is no different than a 2600k minus the ability to change the multiplier. Perhaps with a 3930K you would see your 2x increase on multithreaded applications? Not sure.

Sorry you are disappointed.

Wouldn't the 3930K run even hotter than the 3820?

Maybe I just got a marginal 3820 since it requires so much vcore to
get to 4.8Ghz?

 

[magellan]

So you joined in 2002 and only have 10 posts? That's in line for the lurker of the decade award...

On topic, I would say another benchmark might show a bigger improvement without such a high o/c. Can't really go off one benchmark. Who knows if that is updated to take advantage of newer processors? Not sure

I'm still amazed at the 10 posts in 11 years...

I don't think the post count indicated is correct. Since 2002, the software running
this forum has been COMPLETELY changed at least ONCE. I'll bet my stats were
lost in that shuffle.

But I'm glad you're so easily amazed...

 

[EarthDog]

Wouldn't the 3930K run even hotter than the 3820?

Maybe I just got a marginal 3820 since it requires so much vcore to
get to 4.8Ghz?

Well, we talking heat or temperature as they are two different things. They are both 130W TDP processors. So the heat to dissipate should be the same. Again, loud is a function of the cooler on top of the CPU and what fans you are using on it.

I think that is about right for that 3820... Not many push that thing that far... its usually a 4.4-4.6GHz CPU for the most part. Perhaps you hit the voltage scaling wall...

 

[JoT]

I don't think the post count indicated is correct. Since 2002, the software running
this forum has been COMPLETELY changed at least ONCE. I'll bet my stats were
lost in that shuffle.

But I'm glad you're so easily amazed...

No, they weren't.