A quad with hyperthreading, minimum. Thats who. I'd agree that they'll optimise what they can here and there but if you go back to BF3, the optimisation between beta and final release yielded only around 5% frame rate increase. That's not going to make much difference if your CPU is currently at 100% in the BF4 Beta and you're already bottlenecked.
Well that was fun to fit, the antec 1200 isn't a small case but there is literally only mms to spare space wise the matrix is a beast of a card Runs a lot better now so will see how it is on full release then get saving for a new cpu motherboard and some watercooling Sent from my HTC One X using Tapatalk 2
I can sort of help you get a half decent idea of how your CPU compares if you're up for it.. http://www.3dmark.com/ Download 3dmark (not 11 or PCmark) and run it,and post the link to your results when they pop up. It's about 900mb IIRC and a good way to judge a gaming system (IMO of course, but it seems to be pretty accurate and support everything quite well).
The beta is fine - its Intel fangurls that are crying - remember , AMD threw $8 million at them.... BF3 beta , remember the crying then? not a lot changed between beta and final... except of course the high res added
3DMark isn't anywhere near as cpu intensive as BF4 beta Andy. You can't really compare the two mate. One's synthetic, the other dynamic. I'm not crying. I have 6 cores/12 threads. CPU sits around 60% thank you very much
It doesn't matter if it's synthetic or dynamic so long as it gives you something to compare to. BF4 is a beta and the results so far have been all over the place. With 3Dmark the CPU score tends to be about right scientifically and then you can compare it to newer CPUs that we have results for. That's why I suggested it.. Basically if the CPU Physics score (which is pretty much 100% CPU based) comes back with say, 5000 points then it's obviously quite far behind the 3570k or 8320 @ 4.2ghz (both). That's all I'm trying to do here. Just make a rough prediction As a further example I pitted a G530 Celeron against a E8400 C2D and overall, in everything (and it was quite exhaustive testing) they ended up being around the same
Well thought I may as well run it to see what the ball park figure is without overclocking first result here and results with an oc to 3.8GHz here Physics results in both cases are pretty rubbish
I wouldn't compare physics score to relative in game performance. An FX-8320 gets a physics score of between 8500-9000 depending on the overclock. That's not really that much more than yours; however, the 8320 is very capable and handles BF4 absolutely fine because of the extra cores despite it's low synthetic bench results. This is why a few posts ago, I said that you really cant compare synthetic with dynamic...
I fully appreciate that synthetic tests aren't fully representative of of real life performance but they do give you a ballpark figure. In this case that would be around 55% of the performance of the other cores you mentioned, thread count issues aside, giving my the indication that I best get saving pronto. So far with the beta though swapping my old card for the matrix has made it distinctly more playable with only the odd moments slowing to a crawl and now I can see from one side of the map to the other without just seeing gray blocks Sent from my HTC One X using Tapatalk 2
If you want to see what your pc is doing, download something like HwInfo64 and leave it logging in the background while you run bf4. It records min, max and average for all sensors. Sent from my iPhone using Tapatalk - now Free
In 3Dmark 13 the 8320 clock per clock is around 5% faster (well, 5% more points) than a 3570k at the same clocks. I've not had any concrete Haswell data yet but I would stick my neck on the line and say that clock per clock the 8320 is around the same. However, the 8320 clock per clock has more headroom. What I will say is that my old Xeon (I5 2400) used to score around 6000 points in 3dmark and wasn't enough for any of the new titles. I had trouble with min FPS on - Crysis 3. It was really choking my 670s. Hitman : Absolution. Far Cry 3 (though I did manage 38 FPS min I get over 60 now). Everything else seemed to be a lot more GPU bound. So for example Tomb Raider and Metro : Last Light were fine. I never bothered with BF3 on the Xeon as I'd already completed it with my I7 950 on which it was fine with my 6970 Lightning. The three titles I listed though? CPU cores and MHZ were all that mattered.
You're missing the biggest point though. 3DMark does not take full advantage of multi-core/multi-thread CPUs. BF4 does. 3570K only has 4 cores and isn't multithreaded. 8320/8350 has 8 cores.. which makes a world of difference in BF4... Stop trying to translate 3DMark scores to real-time gaming. It's a benchmark. That is it. A yardstick there to measure the performance of your system against other hardware configurations that are tested with the same benchmark. End of buddy. You can't assume that if your system scores more than another in 3DMark, that it'll be faster than that system in a particular game. Good example was 3770K vs 3930K. On 3DMark performance, the 3930K ran away with the scores but in the real world, there was only a couple of fps difference when set up with the same GPUs. That difference in benchmark scores did not translate to a tangible advantage in gaming fps.
3dmark Firestrike's CPU test uses 8 cores. It also fully supports the FX range of CPUs in Windows 8 64 bit. The results it spits out are pretty much the same results that you will see on a recent game like Crysis 3. I had two of my friends sit and go from 3.8ghz to 4.2 on a 3570k and 2550k and the results were absolutely bang on. I know it's not a game, but, it's a visual benchmark that works in pretty much the same way as a game and puts out results that are completely comparable. If Crysis 3 had a proper benchmark mode I would use that,but it doesn't. It also costs 35 quid. Anything 3770k or 4770k? it's not worth even bothering running them. Simply because they are way ahead of anything AMD offer, even the Centurions. Mind you, they have the price to match.
I'm not disagreeing with you entirely Andy, but it doesn't work in the same way as a game and the results are NOT completely comparable. Well, not to a game anyway, only other 3DMark scores. Benchmarks use complex algorithms to determine a score based on your architecture and hardware. Games don't care about that and the only real measure is fps output. Some favour AMD, some NVidia. It doesn't translate from benchmark to game. I'm not going to fall out about it buddy. If you want to believe that if your system scores x% more in 3DMark than another, therefore it means that it'll be x% faster in a game, be my guest but you're in for a let down. You were right about a proper game benchmark like Crysis 3 though as that isn't synthetic even though it's a fixed benchmark as it's a true measure of your pc in a gaming situation.
Futuremark are using the engine that runs 13 in games. I *think* (though don't quote me on it) there may already be one out there. They were working on a game of their own. As I say, 3dmark scores show the same sort of results as Crysis 3, but it's going to be far more accurate given it does the same thing every time
I give up... EDIT: Actually, I don't give up, not just yet You said:... However, looking at this comparison which is both of the top scores for each CPU, both using a single 7970 (LINK), the 3570K scores 1.91 physics score per Mhz, whereas the 8320 scores a lower score of 1.90 physics point per MHz despite having twice the cores and it ends up with around 1k less 3dmarks overall. Then if you look at another set of #1 results, this time with a single 780 (LINK), the 3570K scores 1.90 per MHz this time around whilst the 8320 scores a much better physics score of 1.96 - but the 3570K scores a massively higher score overall... 3k points in total. Not to be sniffed at. This is proving my point that its not consistent and you simply can't compute relative game performance from a benchmark. And here's the Haswell data you haven't yet seen for BF4... http://www.hardwarepal.com/battlefield-4-cpu-gpu-benchmarks/ The 4670K is outpacing the 8350 at all resolutions, some more so than others. Hope that helps dude.
http://store.steampowered.com/app/223850/?snr=1_620_4__45 £6.45 for the pro version (and firestrike extreme)
