Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
i7 4930 or 4820 for Snowblind engine games?
#1
I'm likely going to be upgrading my computer soon, and looking to get two of my favorite PS2 games to work. Currently i have an i5 2500k, and even overclocked, with the 4k memory patch on PCSX2, running the correct version I can't get BG: Dark Alliance to run properly. At 4.2 GHz, the game will still hit tarry spots where it grinds to a near halt. CoN is much better, but it still has its moments. I will likely be getting either an i7-4930k or an i7-4820k soon (waiting for November sales) and am wanting to know which will help more -- the marginal speed boost from the 4820, or the two extra cores for the 4930?

My other system specs will likely stay the same (unless, the system stability problems I'm having don't resolve themselves).
i7 4930k @4.3, 4x4 GB RAM @2133 (15-15-15-27, quad channel), EVGA 570 @stock, Arch 64b.
Reply

Sponsored links

#2
You are going to have to swap out the motherboard as well considering the I5 2500K is a Socket 1155 and the I7 4820/4930 requires Socket 2011
Reply
#3
Those i7's are pointless if you're just doing gaming and emulation... PCSX2 doesn't use more than two cores by default (three with the MTVU hack) and you get diminishing returns when you set more than 4 or so software threads in Software mode. Your best bet would be to get a Haswell i5-4670k. I think the 4930 is actually Ivy Bridge architecture, which imo is retarded due to Intel's naming scheme.
Reply
#4
Surely you will not need an i7 for emulation but to answer your question:

Number of cores/threads is not important for emulation while frequency is.

Therefore a higher frequency as in i7-4820K would be better. Also the cach per core is probably slightly higher for i7-4820K. But it really depends on the low-level manufacturing.

I would recommend to look for cpu-benchmarks of these cores.

This for example says that the i7-4930K will be better in Single- and Multithreaded applications even if the i7-4930K is only slightly better in Singlecore-activities. If you trust these benchmarks you should buy the i7-4930K.
Reply
#5
There's no point in going for those when he can just buy an i5-4670k and overclock it. Haswell has better single-threaded perf than Ivy Bridge does, and those i7-4930's are Ivy Bridge, which still got beaten in several benchmarks by the Haswell chips because of that.
Reply
#6
For my purposes the Haswell chips aren't that good. I can use the extra PCIe lanes and memory bandwidth, and also the Haswell chips don't overclock as well (so actually the frequency is ultimately lower). For most games you don't need anything more than what I have, and for general gaming, the CPU is an after thought, however these Snowblind Engine games are insanely hard to run for some reasons.

I'm surprised to see that the 4830 was beaten in single threaded benchmarks so frequently by the 4930. Strange. I had seen those before, but didn't look at the 4820 vs. 4930 because I just expected the 4820 to win in single threaded tests. The results from the Haswell CPUs (beating the Ivy Bridge) is almost expected. Since Intel is only competing with itself at this time, they can have different products out on the market that cater to different people without worry of losing in those markets. So its not surprising that Haswell, with its updated architecture, will beat out the Ivy Bridge E processors.

tl;dr: But yeah, the point is that I really can use the socket 2011 CPUs, and that the Snowblind Engine games really do need some beefy processors to play. Also, thanks for the benchmark link and pointing out that 4930 has better single threaded performance. Saw the benchmarks, but didn't pay much attention.
i7 4930k @4.3, 4x4 GB RAM @2133 (15-15-15-27, quad channel), EVGA 570 @stock, Arch 64b.
Reply
#7
Even if you go to one of those processors, the Snowblind games won't really be much easier to play. Ivy Bridge was only about 5% faster than Sandy, and Haswell was 5% over Ivy. The Snowblind games can be lumped in with some of the heaviest games to emulate, which aren't feasible to get fullspeed without high OC's or impossible to currently get fullspeed on at all.

If you need the 2011 chips for something other than PCSX2, then go for it, but you won't see any real benefit whatsoever in PCSX2.
Reply
#8
I see, so brute force won't even make them run better. Thanks.

By the way, what are some of the other hard to emulate games, and also, what would be needed (hypothetically) to run them? I'm not about to get out the liquid helium, but I am curious.
i7 4930k @4.3, 4x4 GB RAM @2133 (15-15-15-27, quad channel), EVGA 570 @stock, Arch 64b.
Reply
#9
Just go for 6Ghz per core and you can emulate everything.

Still pcsx2 can lag since ps2 was lagging for some games and pcsx2 can more or less be only as fast as ps2.
Reply
#10
In 5 years time it wont be an issue for cpu's



(11-05-2013, 12:05 AM)willkuer Wrote: Just go for 6Ghz per core and you can emulate everything.

Still pcsx2 can lag since ps2 was lagging for some games and pcsx2 can more or less be only as fast as ps2.

Come back to reality 6hz on core is not feasble as stable speed
Reply




Users browsing this thread: 1 Guest(s)