Quote:
Originally Posted by MisterFred 
I've been really wanting to see an 8120 Shogun 2 benchmark if you have the game, ever since I looked at the Shogun 2 benchmark on this page:
http://www.bit-tech.net/hardware/cpus/2012/05/01/intel-core-i5-3570k-cpu-review/6
It was noticeable how much ivy bridge blew away sandy bridge there, presumably because of the updated instruction set. Now Bulldozer made a big deal of a supposedly modern & ready for the future instruction set, so it would be interesting to see how it did specifically in Shogun 2. If it has the same instruction benefits, it should at least out pace the phenoms significantly. But Bulldozer had some sort of bug that didn't play well with Shogun 2, so it hasn't been benchmarked. The fix has been out for some time, but bit tech lied about getting around to running a Bulldozer Shogun 2 bench later, and I've wanted to see how Bulldozer runs, particularly since Shogun 2 benches so well on AMD graphics cards (creative assembly had some sort of co-operation with AMD during development). Long story short, it's really the only Bulldozer bench I want to see, particularly if you can come close to the settings the bit tech guys used. Particularly using the CPU benchmark test rather than the more common GPU test.
Edit: The Bulldozer bug is fixed through a bios update of some sort. Apparently it also affects a few other games as well.
I've been really wanting to see an 8120 Shogun 2 benchmark if you have the game, ever since I looked at the Shogun 2 benchmark on this page:
http://www.bit-tech.net/hardware/cpus/2012/05/01/intel-core-i5-3570k-cpu-review/6
It was noticeable how much ivy bridge blew away sandy bridge there, presumably because of the updated instruction set. Now Bulldozer made a big deal of a supposedly modern & ready for the future instruction set, so it would be interesting to see how it did specifically in Shogun 2. If it has the same instruction benefits, it should at least out pace the phenoms significantly. But Bulldozer had some sort of bug that didn't play well with Shogun 2, so it hasn't been benchmarked. The fix has been out for some time, but bit tech lied about getting around to running a Bulldozer Shogun 2 bench later, and I've wanted to see how Bulldozer runs, particularly since Shogun 2 benches so well on AMD graphics cards (creative assembly had some sort of co-operation with AMD during development). Long story short, it's really the only Bulldozer bench I want to see, particularly if you can come close to the settings the bit tech guys used. Particularly using the CPU benchmark test rather than the more common GPU test.
Edit: The Bulldozer bug is fixed through a bios update of some sort. Apparently it also affects a few other games as well.
I don't currently have Shogun 2: total war, but I might buy it if people are interested in how it performs with bulldozer.
The Sabertooth 990fx board that I am using has its BIOS updated to the most current.
Quote:
Originally Posted by Bubba HoteppÂ
Subbed. I'm doing the same thing with an FX-6100 and HD5870 in my 6100 OC'ing thread. Where did you get the mods for Skyrim? It really is a horribly coded game. Why would you offload shadow detail to the CPU and yet only really fully utilize 1 core for the game?? Bad port of a console game IMHO. What area of WOW did you use to bench it?
It's really amazing at the amount of bad information out there in the threads with common sense and reality playing no part at all. When you're looking at 2 CPU's and one is posting minimum 90 FPS in XYZ game and the other is posting minimum 70 FPS, it's not which CPU is better at gaming and which "sucks", it's which one is more overkill than the other (99% of people out there gaming are using 60Hz monitors).
Edit - Which program did you use to log and graph CPU & GPU usage? I've been looking for one like that with no success.
Subbed. I'm doing the same thing with an FX-6100 and HD5870 in my 6100 OC'ing thread. Where did you get the mods for Skyrim? It really is a horribly coded game. Why would you offload shadow detail to the CPU and yet only really fully utilize 1 core for the game?? Bad port of a console game IMHO. What area of WOW did you use to bench it?
It's really amazing at the amount of bad information out there in the threads with common sense and reality playing no part at all. When you're looking at 2 CPU's and one is posting minimum 90 FPS in XYZ game and the other is posting minimum 70 FPS, it's not which CPU is better at gaming and which "sucks", it's which one is more overkill than the other (99% of people out there gaming are using 60Hz monitors).
Edit - Which program did you use to log and graph CPU & GPU usage? I've been looking for one like that with no success.
I got all of my skyrim mods from nexus mods.
I took the WoW benchmark while flying around Azshara, and landing to kill random mobs.
I used aida64 to log the CPU and GPU usage. I ran fraps alongside to capture the framerates, as aida64 doesn't collect that. I then imported those numbers into MS excel and graphed it.
Quote:
Originally Posted by Bubba HoteppÂ
Skyrim is not coded for multiple threads and doesn't use any cpu optimizations (sse2/3, AVX et) and offloads shadow detail rendering onto the CPU. So really you're only going to see 1 core being utilized which is why the 2500K/2600K do better in skyrim because of higher single core performance.
BF3 on the other hand is not CPU intensive at all and GPU performance is far more critical (as is available PCIE bandwidth and GPU mem bandwidth)
http://www.techspot.com/review/458-battlefield-3-performance/page7.html
All you need to play BF3 at max is a 4GHz quad core and a really good GPU.
Edit - I would however expect to see some of those dual cores in the list to dip significantly during a massive multiplayer map.
Skyrim is not coded for multiple threads and doesn't use any cpu optimizations (sse2/3, AVX et) and offloads shadow detail rendering onto the CPU. So really you're only going to see 1 core being utilized which is why the 2500K/2600K do better in skyrim because of higher single core performance.
BF3 on the other hand is not CPU intensive at all and GPU performance is far more critical (as is available PCIE bandwidth and GPU mem bandwidth)
http://www.techspot.com/review/458-battlefield-3-performance/page7.html
All you need to play BF3 at max is a 4GHz quad core and a really good GPU.
Edit - I would however expect to see some of those dual cores in the list to dip significantly during a massive multiplayer map.
I will be running a second round of benchmarks on games like BF3, Crysis2, and MW3 in multiplayer, as well.
Quote:
Originally Posted by Bubba HoteppÂ
Actually, when comparing multi-threaded performance BD is superior to intels Hyperthreading. Basically each "Module" is in essence somewhere between an intel core with HT and a true dual core. The advantage of the BD design is that each "thread" has dedicated hardware resources as opposed to the intel design where each thread has to wait for available resources or utilize resources not currently in use. The reason sandy bridge CPU's do better in some gaming situations is because of the increase IPC of an intel core vs the BD "sub" core (half a module) which makes a difference in games that utilize only one or two cores at most (pointing at skyrim among others). Move to a heavily multi-threaded game and the difference between an 8120/50 and 2500K is not so large anymore and is very close in performance.
Only 2 "cores" are above 40% usage in his graph and he's not using a stock install of skyrim. He's using the modded version. Here's what skyrim unmodified looks like on an FX-6100.
Note the heavy usage on core 6.
Actually, when comparing multi-threaded performance BD is superior to intels Hyperthreading. Basically each "Module" is in essence somewhere between an intel core with HT and a true dual core. The advantage of the BD design is that each "thread" has dedicated hardware resources as opposed to the intel design where each thread has to wait for available resources or utilize resources not currently in use. The reason sandy bridge CPU's do better in some gaming situations is because of the increase IPC of an intel core vs the BD "sub" core (half a module) which makes a difference in games that utilize only one or two cores at most (pointing at skyrim among others). Move to a heavily multi-threaded game and the difference between an 8120/50 and 2500K is not so large anymore and is very close in performance.
Only 2 "cores" are above 40% usage in his graph and he's not using a stock install of skyrim. He's using the modded version. Here's what skyrim unmodified looks like on an FX-6100.
Note the heavy usage on core 6.
My Skyrim install itself is stock, and I am just running mods that do things like increase texture resolutions to 2k, etc. I didn't run anything that changes the way the game runs, though.
Quote:
Thanks. Collecting this information is actually sorta fun.Originally Posted by wint0nicÂ
It's normal, what is not normal is high cpu usage and low gpu usage, nice benchmarks though mate.
It's normal, what is not normal is high cpu usage and low gpu usage, nice benchmarks though mate.
Quote:
Originally Posted by SkipsÂ
Thomas, you are truly and gentleman and a scholar.
Thomas, you are truly and gentleman and a scholar.
Thanks!
Uncensored
(29 items) |
AMD PII Rig
(15 items) |
| CPU | Motherboard | Graphics | RAM |
|---|---|---|---|
| AMD Phenom II 955 Black Edition | ASUS Crosshair iv Formula | SAPPHIRE Radeon HD 6950 2GB (unlocked) - ref | Corsair Vengeance 8GB (2 x 4GB) |
| Hard Drive | Optical Drive | Cooling | OS |
| Seagate Barracuda 1TB 7200RPM | None | NZXT havik | Win 7 Ultimate 64bit |
| Monitor | Keyboard | Power | Case |
| Samsung 19" LCD | Razer BlackWidow Ultimate Mechanical Keyboard | CORSAIR TX650 | Antec 902 |
| Mouse | Audio | ||
| Razer Naga | M-Audio Studiophile AV40 Powered Monitor Speakers | ||
| View all | |||
|
Uncensored
(29 items) |
AMD PII Rig
(15 items) |
| CPU | Motherboard | Graphics | RAM |
|---|---|---|---|
| AMD Phenom II 955 Black Edition | ASUS Crosshair iv Formula | SAPPHIRE Radeon HD 6950 2GB (unlocked) - ref | Corsair Vengeance 8GB (2 x 4GB) |
| Hard Drive | Optical Drive | Cooling | OS |
| Seagate Barracuda 1TB 7200RPM | None | NZXT havik | Win 7 Ultimate 64bit |
| Monitor | Keyboard | Power | Case |
| Samsung 19" LCD | Razer BlackWidow Ultimate Mechanical Keyboard | CORSAIR TX650 | Antec 902 |
| Mouse | Audio | ||
| Razer Naga | M-Audio Studiophile AV40 Powered Monitor Speakers | ||
| View all | |||
