CPU Choices

[SteveGrabowski0]

8 minutes ago, i_build_nanosuits said:

DF tested with an overclocked Titan X @ 1080p...come on...of course in such a CPU starved scenario you WILL see better performance with faster RAM...but for normal users who are bright enough NOT to pair a 100$ dual core CPU with a 1000$ graphics card...normal ram is more than fast enough.

Fast RAM makes a difference in balanced systems too bro. I get a 9.5% better minimum framerate in GTA V and 15% better in Fallout 4 with my DDR3-2400 kit versus my DDR3-1600 kit I was using. It was definitely worth the $80 I spent to replace the kit, so if I was buying from the beginning I would really be happy to only be spending $10-$20 more for the extra performance.

[i_build_nanosuits]

3 minutes ago, SteveGrabowski0 said:

Fast RAM makes a difference in balanced systems too bro. I get a 9.5% better minimum framerate in GTA V and 15% better in Fallout 4 with my DDR3-2400 kit versus my DDR3-1600 kit I was using. It was definitely worth the $80 I spent to replace the kit, so if I was buying from the beginning I would really be happy to only be spending $10-$20 more for the extra performance.

some games benefit from fast RAM i'm aware of that...and if there was only 2 CPU's available (forgetting about the insane upgrade path on skylake) i would probably still pick the FX-8 over the core i3-6100...hands down...it's a more capable processor and it will be relevant longer as games go more and more multi-threaded and GPU focused for performance...but in reality you don't want to lock yourself on such an old socket that does not even have modern commodities such as PCIe 3.0 namely.

And a 250W beast of a CPU like my overclocked FX was, it trow up a shit ton of heat in the case and if you also have an high-end or powerhungry GPU to top it off...you see what i mean i'm sure.

[Prysin]

3 minutes ago, SteveGrabowski0 said:

Fast RAM makes a difference in balanced systems too bro. I get a 9.5% better minimum framerate in GTA V and 15% better in Fallout 4 with my DDR3-2400 kit versus my DDR3-1600 kit I was using. It was definitely worth the $80 I spent to replace the kit, so if I was buying from the beginning I would really be happy to only be spending $10-$20 more for the extra performance.

indeed. After all, you got no downside to buying faster RAM (if your mobo supports it) other then price....

[Prysin]

33 minutes ago, i_build_nanosuits said:

some games benefit from fast RAM i'm aware of that...and if there was only 2 CPU's available (forgetting about the insane upgrade path on skylake) i would probably still pick the FX-8 over the core i3-6100...hands down...it's a more capable processor and it will be relevant longer as games go more and more multi-threaded and GPU focused for performance...but in reality you don't want to lock yourself on such an old socket that does not even have modern commodities such as PCIe 3.0 namely.

And a 250W beast of a CPU like my overclocked FX was, it trow up a shit ton of heat in the case and if you also have an high-end or powerhungry GPU to top it off...you see what i mean i'm sure.

Although, to be fair, no consumer GPUs, even AMDs R9 390Xs in quadfire will be able to fully saturate PCIe Gen2.

AMD has validated all their "Crossfire over PCIe" solutions to run without any bottleneck at PCIe Gen2 x8.

 

Hell, some of the "new" 990FX boards has M.2 slots with 4x PCIe Gen2... sure it may not beat intel, but i do not think you will need more then 1500MB/s read/write speeds anytime soon

 

That being said. The downside of AM3+ is perhaps not the longevity of the boards, or the parts. but rather the inconsistent quality. You REALLY got to know what you are doing when buying, or you can end up with a fire situation (MSI Krait)

 

Oh and fun fact: FX 8320, 4.62GHz at 1.47v -> Noctua NH-D9L 92mm cooler w/2 fans.... tops out around 59c. Sure it is waaaay too hot. But it is DEAD SILENT, even while the mobo is sitting openly right next to me. A Hyper TX3 for 15 bucks should do 4.3-4.5GHz at 1.42-1.44v easily.

 

EDIT:

This AM3+ board is damn good.

http://pcpartpicker.com/part/gigabyte-motherboard-ga990fxgaming

 

They cut back on the insane USB2.0 lanes (chipset offers something like 14 USB2.0 connections) the 990FX chipset uses and merged them into USB3.0 i think.... thats how they somehow managed to get USB 3.1 on there.

 

https://en.wikipedia.org/wiki/AMD_900_chipset_series

 

AMD 990FX gives you 38 PCIe Gen2 lanes for various connectors + 4PCIe Gen2 lanes internally for use in the Northbridge.... so total of 42 PCIe Gen2 lanes. Which would be equal to 21 PCIe Gen3 lanes... Meaning total bandwidth wise, these boards arent that far behind intels consumer chipsets.

[Sintezza]

4 hours ago, i_build_nanosuits said:

DF tested with an overclocked Titan X @ 1080p...come on...of course in such a CPU starved scenario you WILL see better performance with faster RAM...but for normal users who are bright enough NOT to pair a 100$ dual-core CPU with an overclocked top-end 1000$ graphics card...normal ram is more than fast enough.

 

Exactly that.

 

DF just trying to start a Bullshit hype.

 

 

[Sintezza]

4 hours ago, SteveGrabowski0 said:

Fast RAM makes a difference in balanced systems too bro. I get a 9.5% better minimum framerate in GTA V and 15% better in Fallout 4 with my DDR3-2400 kit versus my DDR3-1600 kit I was using. It was definitely worth the $80 I spent to replace the kit, so if I was buying from the beginning I would really be happy to only be spending $10-$20 more for the extra performance.

 

I dont believe this realy.

What are your system specs?

Do you have any documentation for backing it up?

 

Because i have exaly done similar tests with DDR3-1600 VS 2400, in a balanced system.

And in most games that i have tested, there was no significant diffrence.

There might be a very few games that slightly benefit from it in certain scenario´s.

But till now i havent realy seen it.

So if someone has any reputable sources for this let me know.

 

Because i´m very interested to see those.

Digital Foundry tests make no sense to me, for the same reason @i_build_nanosuits above also mentioned.

 

I also think that Linus has tested it aswell.

[SteveGrabowski0]

21 minutes ago, Sintezza said:

 

I dont believe a single shit of this.

What are your system specs?

Do you have any documentation for backing it up?

 

Because i have exaly done similar tests with DDR3-1600 VS 2400, in a balanced system.

And in most games that i have tested, there was no significant diffrence between ram speeds.

My system specs are in my sig: Xeon E3-1231v3 + GTX 970 at 1080p. What do you mean documentation? I didn't take video of it if that's what you mean. @MageTank reports similar results. Digital Foundry reported similar results in one of their Fallout 4 articles when using a 4690k and a GTX 970. I think the article was titled something like 'What does it take to run Fallout 4 at 1080p 60 fps'. I would never say it makes a difference in all games. It doesn't make a lick of difference if I play a strongly gpu bound game like Shadow of Mordor. But it was obviously noticeable in GTA V. There is an area near Mt Chiliad that brings my fps to the lowest I can get them in normal gameplay at mostly ultra settings (I'm not counting doing donuts in the grass and kicking up all kinds of dirt and crap which tanks framerate). With my DDR3-1600 kit that was a guaranteed drop to 52 fps. With my DDR3-2400 kit the drop is brought up to 57 fps, and it's also much rarer. Now I can only get that low of a drop if it's near sunset and the sun is low and casting pretty complex shadows. Whereas I could always get that drop to 52 fps on the DDR3-1600 kit. 5 fps may not sound like a lot, but there is a huge difference in fluidity when you're targeting 60 fps. In Fallout 4 my worst framerates came inside buildings with light coming through windows and lots of particles floating around, such as inside the speakeasy in Concord. I'd drop to 48-51 fps there with my DDR3-1600 kit. Exact same system except with my DDR3-2400 kit and I'd drop to 57-58 fps there. This is with everything at ultra except godrays and shadow draw distance, which I put at medium since they hammer cpus and I'm running a locked Xeon. This is also before the patch that gave all the particle effects from shooting walls and stuff though, I haven't played FO4 since that 1.3 patch came out.

[SteveGrabowski0]

23 minutes ago, Sintezza said:

I also think that Linus has tested it aswell.

Linus' test is ridiculous and no one should pay any attention to it. It's just a gpu benchmark since he is running games at really high AA settings with a weak gpu: a GTX 660 Ti. My system with a 970 would be strongly gpu bound at 1080p with 8xMSAA on Far Cry 3, with a 660 Ti it's a nightmare. Same with 4xSSAA on Metro Last Light. SSAA is so punishing on a gpu.

 

 

[Sintezza]

I have seen Megatanks tests yes, but again a dual core with a GTX770.

 

if you have balanced cpu + gpu combination, faster ram speeds will make no significant diffrence in most games.

There might be a very few games that slightly benefit from it, but till now i have not seen it.

Biggest improvement i have seen was arround 3 fps on average in Dirt 3.

I cant speak for FO4 i might test them wenn i have done my personal system upgrade.

 

I do agree that faster ram isnt that much more expensive.

But till now i haven't seen much real life benefits from having it.

I might do some new tests as soon as i have finnished my upgrade.

I might make a video or write an article on it.

 

[Nena Trinity]

6 hours ago, i_build_nanosuits said:

yes...but the athlon is cheaper and that's why it make a decent option for super budget rigs...FM2+ boards are dirt cheap as well.

Going with the FX-6300 and a decent overclocking AM3+ motherboard for it (and most likely an aftermarket cooler) put you again in a spot where you could probably sneak an i3-6100 and a cheap H series motherboard for around the same price...and this i3 in games is better than even your 8370.

 Nope at the best the i3 is on par unless the game only uses 2-cores then it win for sure. Not to mention none-gaming stuff then the i3 falls further behind, but if gaming only for noobs the i3 should be easier to setup because all they have to do is set XMP to on!  

[Dabombinable]

4 minutes ago, Sintezza said:

I have seen Megatanks tests yes, but again a dual core with a GTX770.

 

if you have balanced cpu + gpu combination, faster ram speeds will make no significant diffrence in most games.

There might be a very few games that slightly benefit from it, but till now i have not seen it.

Biggest improvement i have seen was arround 3 fps on average in Dirt 3.

I cant speak for FO4 i might test them wenn i have done my personal system upgrade.

 

I do agree that faster ram isnt that much more expensive.

But till now i haven't seen much real life benefits from having it.

I might do some new tests as soon as i have finnished my upgrade.

I might make a video or write an article on it.

 

That's what I found with my Xeon X5450 when I was running it with my GTX 970, actually overclocking the RAM affected the FPS more than overclocking the Xeon-though in Cinebench at around 4.4-4.6GHz it would match an i5 4440-which don't bottleneck GTX 970. With my GTX 650ti however (given it+my Xeon to a friend), no matter how much I overclocked the RAM, it didn't have as large an impact as overclocking the CPU-and then the impact of overclocking the CPU was only felt in games where the CPU was maxed out.

[SteveGrabowski0]

11 minutes ago, Sintezza said:

I have seen Megatanks tests yes, but again a dual core with a GTX770.

 

if you have balanced cpu + gpu combination, faster ram speeds will make no significant diffrence in most games.

There might be a very few games that slightly benefit from it, but till now i have not seen it.

Biggest improvement i have seen was arround 3 fps on average in Dirt 3.

I cant speak for FO4 i might test them wenn i have done my personal system upgrade.

 

I do agree that faster ram isnt that much more expensive.

But till now i haven't seen much real life benefits from having it.

I might do some new tests as soon as i have finnished my upgrade.

I might make a video or write an article on it.

 

I really don't see the problem with Digital Foundry's tests. The i3 + Titan X is wildly unbalanced, but they show nice gains in some games with an i5/i7 and Titan X also. And that's not a ridiculous combination considering the upcoming GP104 and Polaris 11 cards should outperform the Titan X, based on past history with node shrinks. I haven't played Witcher 3 since getting my DDR3-2400 kit and even if I had I never took note of what my worst performance was there with my DDR3-1600 kit, so I can't back up their results showing a big gain in Novigrad from faster ram though.

[i_build_nanosuits]

17 minutes ago, SteveGrabowski0 said:

I really don't see the problem with Digital Foundry's tests. The i3 + Titan X is wildly unbalanced, but they show nice gains in some games with an i5/i7 and Titan X also.

that is also true though, i have to give you that.

 

i was not really questioning the fact that faster ram can make a difference in games earlier...what i was more against is the whole ''you also need faster memory with the core i3-6100 in order to get good performance out of it which makes the cost go up'' which honestly is not really a factor when comparing an FX-6300 or FX-8320 with the i3-6100 in games because even with the cheapest barebone green PCB DDR4 kits it will still outperform both of these in most games. that's all didn't really wanted to start a debate on faster memory help performance in games...i think it does, to some extent, depending on the game, settings and GPU you use

sorry about that

[Sintezza]

24 minutes ago, SteveGrabowski0 said:

I really don't see the problem with Digital Foundry's tests. The i3 + Titan X is wildly unbalanced, but they show nice gains in some games with an i5/i7 and Titan X also. And that's not a ridiculous combination considering the upcoming GP104 and Polaris 11 cards should outperform the Titan X, based on past history with node shrinks. I haven't played Witcher 3 since getting my DDR3-2400 kit and even if I had I never took note of what my worst performance was there with my DDR3-1600 kit, so I can't back up their results showing a big gain in Novigrad from faster ram though.

I will look if i can find more information about which particular games might benefit from it.

Would be nice if there was an overview on it, so it can be tested.

Of course i dont say that you are wrong by any means.

Its just that i personaly have only see a slight advantage in 2 games till now.

Dirt3 and Grid autosport.

 

I dont realy think that there are manny games that will benefit from in the total picture, unless the cpu hitting its limmits ofc, or the GPU hits its ram vram limmits

Which is what you see with digital foundry tests pairing a TitanX OC with a dualcore chip on 1080p.

 

If Digital foundry would do the same tests with lets say a i5-6600K + TitanX at 4K, then the ram speed should basicly again make no significant diffrence.

BUt thats because you are mainly gpu boand then.

[anterioreight27]

yea i agree i love amd and have a amd cpu get a  i3 6100  if upgrading