The TWC Total War Benchmark thread - compare your real TW performance

[mennelik]

would a fx 6300 vishera bottleneck a GTX 700 series card

[alQamar]

Hard to say Mennelik, all existing consumer processors bottleneck GPUs in Rome 2, so at first I would answer yes but I have to less data to estaminate how much. Also the 700 series has a wide spread from a future 750TI to 780TI. Which one you are looking for? The most reasonable buy would be a 760 Ti.

[SamueleD]

Hard to say Mennelik, all existing consumer processors bottleneck GPUs in Rome 2, so at first I would answer yes but I have to less data to estaminate how much. Also the 700 series has a wide spread from a future 750TI to 780TI. Which one you are looking for? The most reasonable buy would be a 760 Ti.

You mean 760? There is no 760 Ti!
I think that, based on the various benchmarks we ran both on Rome II and Shogun 2, and the various benchmarks I saw elsewhere an FX 6300 is likely to bottleneck TW performance (with pretty much any current 700 series) a bit more than a comparatively priced current Intel CPU (an i3). Granted, it's a very rough estimate, with no direct evidence (no 4th gen i3 vs 6300 comparison available). As for other games, unless they are as heavily CPU bottlenecked as TW, they are likely not to be bottlenecked by an FX 6300.
But I digress, as you say any CPU will bottleneck pretty much any GPU in Rome II, so to answer mennelik in Total War, yes, it would, like any other CPU, unfortunately.

[alQamar]

You mean 760? There is no 760 Ti! I think that, based on the various benchmarks we ran both on Rome II and Shogun 2, and the various benchmarks I saw elsewhere an FX 6300 is likely to bottleneck TW performance (with pretty much any current 700 series) a bit more than a comparatively priced current Intel CPU (an i3). Granted, it's a very rough estimate, with no direct evidence (no 4th gen i3 vs 6300 comparison available). As for other games, unless they are as heavily CPU bottlenecked as TW, they are likely not to be bottlenecked by an FX 6300.

Well someone made benchmark with Rome 2 on release day and the poor Vishera performance got a lot better in Rome 2 than it was in Shogun 2. As said it is very complicate to give any hardware advices for Rome 2.

[SamueleD]

Well someone made benchmark with Rome 2 on release day and the poor Vishera performance got a lot better in Rome 2 than it was in Shogun 2. As said it is very complicate to give any hardware advices for Rome 2.

It did? Good for AMD! Once the patch-storm ends, and proper drivers are finally out (for CFX and SLI too), we should get to benchmark this game and put new life into this thread, it seems (from looking around with PFM) that it would even be possible to create custom benchmark scenarios!

[alQamar]

well we will see. Personally I am of course interested into it but the auto degration and missing log files will not be able to produce enough data for me to verify things. I really do not get why the CA decided to kill the log files? What's wrong with them.

[mennelik]

tnx for the info guy's but now I have a evga sc 650 ti boost with a extra oc and still not enough juice to run RTW2, but I have to mention with a modded unit size 240 and some graphic tweaks, that's why I would like to upgrade in the future to a 700 series card so that I can level up the fps to enjoy big battles and with SH2 no problem at all( only ultra)
just to be sure

hopefully patch 9 will have a big impact on performance

[alQamar]

well the 650 ti boost isn't a very speedy card for Rome 2. Instead of feeding the performance troll of Rome 2 with countless money, you should be patient for future patches as you suggested yourself.

[A Barbarian]

tnx for the info guy's but now I have a evga sc 650 ti boost with a extra oc and still not enough juice to run RTW2, but I have to mention with a modded unit size 240 and some graphic tweaks, that's why I would like to upgrade in the future to a 700 series card so that I can level up the fps to enjoy big battles and with SH2 no problem at all( only ultra) just to be sure

If u increase unit size, you will just create a bigger CPU bottleneck. In that case updating to a 700 series card will not help u one bit for big battles.

hopefully patch 9 will have a big impact on performance

Sorry, it won't. No patch ever will increase the performance of Rome 2 much. Trust me!

[alQamar]

While you are essential true Barb, I give them a last chance to fix this till patch 11. They talked about a "big" patch 3 patches ahead of 8.1. We'll see.

[A Barbarian]

No patch ever will increase the performance of Rome 2 much. Trust me!

It seems that I have to eat my words since patch 9 actually improved the minimum framerates in battles when zoomed in a lot.

But the funny thing about is, that it seems that they have reverted code to Shogun 2 routines. The CPU usage pattern (regarding utilization of multiple cores) looks much more like in Shogun 2 rather than in patch 8. Fewer CPU cores are used than before, just like in Shogun 2.

The fact that performance increased despite the usage of fewer CPU cores is no contradiction when you know about multi-threading programming. Different people and myself (I am a programmer) have pointed out that multi-threaded code can be slower than single threaded code even though it only uses 1 CPU core insted of 2. That's because data synchronization between different threads is complicated and very costly. And if you can't distribute your workload efficiently with few synchronizations then multiple threads will be slower.

[Blarni]

Anyone bought a lower-middling price laptop recently that plays RTW2 well? Doesn't need to be ultra.

PS:
I know this question would have been asked/answered a number of times in these forums - but when I do see it, it tends to be pre-release and very hypothetical.

PPS: I know desktops are better. I just need a laptop.

Thnx

[alQamar]

Anyone bought a lower-middling price laptop recently that plays RTW2 well? Doesn't need to be ultra.

PS:
I know this question would have been asked/answered a number of times in these forums - but when I do see it, it tends to be pre-release and very hypothetical.

PPS: I know desktops are better. I just need a laptop.

Thnx

Well this is the wrong thread for that request:
here is the appropriate one: http://www.twcenter.net/forums/showt...-Advice-Thread

[Blarni]

Well this is the wrong thread for that request:
here is the appropriate one: http://www.twcenter.net/forums/showt...-Advice-Thread

Ah, thanks

[AstroCat]

Here are my results with Rome 2 - patch 9.

Full - 49.4fps
Extreme - 55.2fps
Ultra - 88.6fps
High -188.2fps

Basic Specs:
Processor (CPU): i7-2600k@4.2GHz
Motherboard (Mobo): Asus P8Z68-V Pro
Graphics Card (GPU): EVGA GTX Titan
System Memory (RAM): 8GB (2 x 4GB) G.Skill Ripjaws X Series
LCD: Asus VE276Q 27" (1920x1080)
Soundcard: Creative Labs X-Fi XtremeGamer Fatal1ty Pro
OS: Win7 Pro 64bit

Drivers (at default for the test) 334.67-beta
Ambient Occlusion: Not supported for this application
Anisotropic filtering: Application-controlled
Antialiasing - FXAA: Off
Antialiasing - Gamma Correction: On
Antialiasing - Mode: Application-controlled
Antialiasing - Setting: Application-controlled
Antialiasing - Transparency: Off
CUDA - GPUs: All
Double Precision: None
Maximum pre-rendered frames: Use the 3D application settings
Multi-display/Mixed-GPU Acceleration: Multiple display performance mode
Power Management Mode: Adaptive
Texture Filtering - Anisotropic Sample Optimization: Off
Texture Filtering - Negative LOD Bias: Allow
Texture Filtering - Quality: Quality
Texture Filtering - Trilinear Optimization: On
Threaded Optimization: Auto
Triple Buffering: Off
Vertical Sync: Use the 3D application settings

[alQamar]

Thanks that a good base for a comparison to the next patches

[Biggus Splenus]

It'd be nice to see the CPU core voltages that people had for each benchmark... might be fun to compare how lucky everyone was in their silicon quality

4 core CPUs with hyperthreading technology:
Intel Core i7-4770K, 4,5 GHz, Haswell C1, 22nm lithography, 84 Watts TDP, DDR3 2666 MHz CL11-13-13-35, Dual Channel: 52 FPS

:O !!!!!!!!!
A clear winner; that's amazing running memory at 2666Mhz AND CPU at 4.5Ghz. Someone was lucky with their purchase

EDIT: I realised I haven't visited this place in a long time, so I thought I might. After what seems to be a dozen motherboard BIOS updates, it seems that use to be the problem with my system. I remember me an alQ trying so hard to figure out the problem, so there you go . Managed to get 43.1fps on my 4670K @4.4Ghz + 1600Mhz CL9 ram. This seems to be pretty in line with everyone else's results, and I think it's just the average ram speed that is slowing my benchmark down I might try rise it to 1866Mhz, but it seems most RAM these days can only run what they're rated at (mine is a 1600Mhz kit, so that's probably the limit).

Rising it to 4.5Ghz didn't give much improvement at all. Rising my cache speed to the same speed as my CPU seemed to improve performance too, only by about 1fps though. Couldn't get cache speed up to 4.5Ghz, and seems as getting the CPU to 4.5Ghz didn't give much more performance (and it made the package run at 80 degree celsius @ full load), I just decided to keep them both @ 4.4Ghz, which seems to work well. I hate noisy PCs, so that temp really annoyed me that's why I pack nothing but Noctua when I can.

This is Shogun 2 benchmark by the way.

EDIT: I also ran the GPU benchmark with a few variables... tried it with a temp limit of 70 celsius (82% TDP), 80 celsius (100% TDP), and 95 celsius (106% TDP)... the latter was the maximum limit, and it seems to have only taken that as a suggestion... it really only got to 82 celsius Anyway, I got 88fps, 91fps, and 92fps respectively.

EDIT AGAIN: All the GPU benchmarks are with reference speeds, I'll try overclocking soon.

[SamueleD]

At 1.25V I can get my 4670K @4.4GHz and 4.3GHz cache (cache voltage 1.25V), 4.4GHz cache was not completely stable, to get to 4.5GHz it needs a lot more voltage, and I'd rather keep my temps as they are, RAM is 1600MHz, CL8, I guess I was unlucky, no 4.6GHz - 4.8GHz for me. I'll run the benchmarks again, when I have the time.

[Biggus Splenus]

I've got 1.22v @ 4.4Ghz on core and cache, so I guess I was just a little bit more lucky. Takes 1.25v to get the core to 4.5, but the cache wont go any higher, no matter what I try. I'd also prefer my chip running a bit cooler, so I keep it at 4.4Ghz... I would try pushing it to 4.6 or 4.7, but my cooling isn't that insane in performance.

EDIT: increased ram to 1866Mhz while keeping timings at CL9, and I got 44.2fps

2133 @ CL10 gives 45.7fps... THAT'S A NEW RECORD FOR 4670K ON HERE! And even rivals the 3930K o.O but actually that's not surprising

[Mietek88]

My results with Rome 2 - patch 9.1

Extreme - 37.4fps
Ultra - 57.4fps
V. High -74.1fps
High -108.5fps

Driver - last beta 14.1

Basic Specs:
Processor (CPU): i5-750@4.0GHz
Motherboard (Mobo): Asus P7P55D
Graphics Card (GPU): Gigabyte 280x
System Memory (RAM): 8GB (2 x 4GB) Corsair Vengeance 1600LCD: HP LP2475w 24" (1920x1080)
OS: Win7 Pro 64bit