Black Hole - Benchmark (OLD Version)

Status
Not open for further replies.
This bench certainly seems a lot more stable than the last one. Ran the bench again tonight and got 1348 again.
 
This bench certainly seems a lot more stable than the last one. Ran the bench again tonight and got 1348 again.

yeh, but I have to scale every test the same. That's what I gonna do in a month. It's not all this bench, it's just amd that gets beaten in floating point.
 
score update:

benchmarktablev3.jpg
 
yeh, you have to look to the singlethreaded. The phenom is beaten there...
That's a big point were amd has to work on.

It's still not accurate. Clock for clock the Phenom II is around the same as core 2 architecture, with some core 2 being a little faster BUT not a pentium dual core which is stripped of it's cache. Now take that into account, and then take into account my Phenom II was almost 600 MHZ faster, there's absolutely no way that a pentium dual core at 2.9ghz is faster than a Phenom II at 3.5ghz.

Now here's a semi accurate example, E6300 VS a 955-

http://www.anandtech.com/bench/Product/94?vs=88

Thats only a 400mhz speed difference, my Phenom II was nearly 600mhz faster....yet even at only 400mhz faster the Phenom II pulls way ahead over the pentium dual core in single threaded benches.

There's simply no way a Pentium dual core at 2.9ghz should be even close as fast as Phenom II at 3.5ghz, let alone beat the Phenom II thats just nuts.


Edit- Take a look, a Q8300 at stock 2.5ghz is pulling a 720 and my Phenom II at 3.5ghz is at 734. Absolutely no way. Q8xxx is around the exact same speed clock for clock as Phenom II, but I was 1 entire ghz faster! Q9xxx's are slightly faster than Phenom II but nothing ground breaking like nehalem and sandy bridge. So, in the real world, my Phenom II at 3.5ghz would be ~25% faster than a Q8300 at 2.5ghz, so roughly it should be around 900 points on your bench. And that sounds about right since a Q6600 at 3.2ghz is scoring 877 points, and Phenom II is easily as fast as kentsfield core 2 architecture.
 
Last edited:
Plz read what I posted above, amd is scoring extremely bad at the floating point maths. In the next version I gonna fix that part.
If you want I can fix it now. But there are already too much tests to begin again while it's just released.
 
haha OK well I wasn't following what was posted after I only replied to your next post after my last one.
 
87dtna said:
Q9xxx's are slightly faster than Phenom II
Depends which Phenom II and which Q9xxx chip you're talking about. I make the 965 BE to be roughly the same speed as the Q9650, but if you compare the 965 BE to the Q9300, the 965 BE is the clear winner. In some cases the 965 BE is quite a bit faster than the Q9300.
 
Depends which Phenom II and which Q9xxx chip you're talking about. I make the 965 BE to be roughly the same speed as the Q9650, but if you compare the 965 BE to the Q9300, the 965 BE is the clear winner. In some cases the 965 BE is quite a bit faster than the Q9300.

Clock for clock man, don't take my quote out of context. Phenom II is roughly the same as Q8xxx clock for clock, Q9xxx's are all faster. Q9x5x's are even faster yet.
 
Anybody here got an E8400 or something they can try this on? I want to see how it does compared to the rest of the processors here.
 
The only bench that seems truly un biased AMD VS intel I've found is Wprime.

It all depends on the floating point maths. But I've to include it, because it's a part of the calculations a cpu must do. Without that a phenom would be almost the same as an i7 920...
 
Anybody here got an E8400 or something they can try this on? I want to see how it does compared to the rest of the processors here.



It would be slightly faster than the E6500, all it has is more cache it's virtually the same CPU. I would ball park that it would come up around 675.
 
Last edited:
It all depends on the floating point maths. But I've to include it, because it's a part of the calculations a cpu must do. Without that a phenom would be almost the same as an i7 920...

A phenom II at 3.5ghz should actually be pretty much the same as a stock I7 920 in the single threaded bench while the I7 would still win in multithreaded because of HT. So a stock 920 scored a 923 on your bench, this all still makes sense that a Phenom II at 3.5ghz should score close to 900.
 
Last edited:
A phenom II at 3.5ghz should actually be pretty much the same as a stock I7 920 in the single threaded bench while the I7 would still win in multithreaded because of HT. So a stock 920 scored a 923 on your bench, this all still makes sense that a Phenom II at 3.5ghz should score close to 900.

Ok, I'll scale the floating point maths lower then the integer maths.
Pfff, I want to release it again lol. I released it too early :'(
 
How does an i7 870 beat an i7 920 when both are at stock speed?

Lynnfield is faster than bloomfield, this is mainly due to a faster uncore speed (The "uncore" refers to PCIe controller, memory controller, DMI/QPI and the L3 cache). Uncore runs at 2.40GHz on the 870 and only 2.13GHz on the 920. But thats not the only speed differences because an I5 750 is actually faster than an I7 920 in single threaded apps and the I5 has 2.13ghz uncore. So Intel had to have done something to increase performance. Lynnfield is newer than bloomfield, perhaps they improved something on the architecture.

But with that said, the 870 is clocked at 2.93ghz and the 920 is at 2.66ghz stock.

i7 920 is quad like i7 870. i7 920 isn't really fastest as i7 870 since it is low class for LGA 1336.

Thanks for your worthless post, you've said nothing that makes any sense nor has any truth to it.
 
Last edited:
Status
Not open for further replies.
Back
Top