Computer Forum Chat Thread

[Xuvial]

1080p is about as high as I would go for CPU tests (even with CFX/SLI), 1440p has just got "extreme GPU bottleneck" written all over it.

[Faithh]

1080p is about as high as I would go for CPU tests (even with CFX/SLI), 1440p has just got "extreme GPU bottleneck" written all over it.

Well testing cpu's when you're GPU bound everywhere comes to the same logic as testing the new R9 290x vs Titan with a Pentium 4 1.6GHz since that cpu is obviously a bottleneck, imo. I dont know why Anandtech does this kind of benchmarks since they're very respected as a source. Just to tell people that a certain CPU is capable to play a game?

[Deleted]

tommorow the last piece of the watercooling puzzle will finally arrive! (assuming ups doesn't show up when i'm at school)
also as usual , murphy is right regarding timing.
the week when i have to do shitloads of schoolwork the worst distraction arrives.

[Butler to Baby Sloths]

Well testing cpu's when you're GPU bound everywhere comes to the same logic as testing the new R9 290x vs Titan with a Pentium 4 1.6GHz since that cpu is obviously a bottleneck, imo. I dont know why Anandtech does this kind of benchmarks since they're very respected as a source. Just to tell people that a certain CPU is capable to play a game?

Perhaps to dispel peoples' misconceptions about getting an i7 extreme for a gaming rig, what with places like Alienware, iBuyPower et al pushing 3820s and 3930s in their top level "gaming" computers?

[Majesticii]

It's not a misconception, that's the point. A 4670K IS going to give you a better experience overall compared to a FX-8350. It's just annoying that this type of benchmarking only demonstrates the GPU's capability, not the CPU. This benchmark would have people believe a 3930K gives you roughly the same fps as a 5800K. Which is ofcourse, rediculous.

[BicycleMafioso]

It's not a misconception, that's the point. A 4670K IS going to give you a better experience overall compared to a FX-8350. It's just annoying that this type of benchmarking only demonstrates the GPU's capability, not the CPU. This benchmark would have people believe a 3930K gives you roughly the same fps as a 5800K. Which is ofcourse, rediculous.

I like that benchmark. It shows how skewed people are toward Intel for gaming.
How is this testing the GPU? What they say in the graph is "CPU doesn't matter". I would agree.
Why is it ridiculous? FPS is FPS. That's what people wants to know.

Outside of Wow and heavily graphical RTS, Intel has no inherent benefit over AMD. CPU is growing less and less important in gaming. Which is great. And AMD's CPUs is more than good enough. Are they weaker CPUs? Sure. But so is 4670K compared to 4960X. It's still no point in going for the 4960x if you intend to game, and the same could be said between FX 8320 and i5 4670K; You can't apply the same logic once and then disregard it in the same identical occurence. the i5 is stronger than the FX, yes. Does it matter for most gaming? No. Is it thus a "waste" spending that extra money on the i5 over the FX if your primary intent is gaming? Definitely, yes.
Then you can disqualify that by saying that you like intel's better, and that you want a stronger CPU and that is fine*. But it's the exact same scenario and you still draw two different conclusions.

(This is not directed at anyone specific, I'm just thinking aloud.)
* = I chose the i7 4770K, I guess that should tell you about my preference. I don't intend to lie to myself though.

[Xuvial]

the i5 is stronger than the FX, yes. Does it matter for most gaming? No. Is it thus a "waste" spending that extra money on the i5 over the FX if your primary intent is gaming? Definitely, yes.

"Most gaming" = depends on which games the person is playing = decides whether the 4670K is worth it for them or not.

[Faithh]

Perhaps to dispel peoples' misconceptions about getting an i7 extreme for a gaming rig, what with places like Alienware, iBuyPower et al pushing 3820s and 3930s in their top level "gaming" computers?

Personally I think that those graphs are just more for people who actually know something of hardware or can build a PC. Many ppl who plan to buy an alienware wouldn't even know there is a cheaper platform like 1155. Doubt they'll ever do some google but rather take their credit card out and buy it immediately.

[DeltrusDisc]

Went to a beard festival last night!

The "Inaugural Chicago Beardsmen Oktobeard Fest!!!" It was charitable to raising money for a bearded brethren's medical issues and was hosted at a place that servers SUPERBLY RARE AND AMAZING CRAFT BEERS! How could it get better? By having people compete with their beards!

Thus, I thought I'd share a couple pictures with you all, my closest computer forum nerdy brethren from around the world! Cheers!!

The general crowd:


Got third place in the Natural Goatee section!


The fine gentlemen I lost to!


Head on over to this Facebook page for pictures of all the fine beards and beardettes!

https://www.facebook.com/media/set/?...3088180&type=3

[BicycleMafioso]

Hey.
Hey-hey-hey.

You can't just go and post that here. It's wildly off-topic. You get right back here and review that beer!
You can't mention Oktobeardfest and leave out the beer. There was beer, right? Else I'm thoroughly disappointed.

(else, I have no issues with the post)

[Majesticii]

How is this testing the GPU? What they say in the graph is "CPU doesn't matter". I would agree.

But it does matter, it only takes a CPU-bound game to show it. You dissappoint me, Tetris...



I rekon a 4670K beating the FX8350 by roughly double the framerate. Which would be VERY important for a 120HZ setup. Were it not for this game in the whole review, you would not come to this same conclusion. In a CPU test, you want to see the CPU's influence or performance in today's popular games. You want to see if it can reach that 120fps, or has sufficient performance to make use of that expensive CF/SLI setup. To not see your framerate stay the same after upgrading to two graphics cards, wondering what is wrong. If i want to see how well Sleeping dogs scales on CF, or how much framerate a HD7970 reaches at 1440p max settings i'd watch a HD7970 review..

[BicycleMafioso]

But you linked 1440p, so 60 Hz. Anything above is superfluous.
EDIT: And for the matter, I was thinking of Civ5 when I said "RTS", even though it is turn-based. There I stupided. Otherwise, I think that's a perfect example of the point I was trying to make; CPU doesn't matter.

[Xuvial]

Depending on the game it sort of matters, but yeah it's not a huge amount.

http://www.bit-tech.net/hardware/201...l-cpu-review/5

[Majesticii]

But this still doesn't give you ANY indication on the performance difference whatsoever. Why would you not be interested in performance differences or achieveable framerates with different processors. This way of thinking, and benching, is seriously narrowminded. You have no guarantee that you will be able to reach 60fps in newer, more demanding games released shortly after. Because you have no clue on relative performance. Look at the X4 series struggling in BF4 now:



When the (then) same priced i5-760 still achieves a reasonable framerate. You invested the same, but the AMD system is already unplayable. Relevant data, is relevant. You will see this same difference in about a year, between the 3570K/4670K and the FX-8350. When they cost relatively the same now.

I seriously despise the method of CPU testing these days. Because you cannot see the difference properly. Honestly, i think they do this to create a market for both processors. Because when the FX series goes toe-to-toe with the i5's on low settings, it's shameful just how bad they are.

[BicycleMafioso]

But this still doesn't give you ANY indication on the performance difference whatsoever.

Well, they kind of do, in the regard that there is no performance difference under a scenario such as the one in the test. "Relative performance"? In that scenario, none. Reaching 60 FPS in a newer game? No, not even some of the best hardware out there today will do that. In say, Star Citizen.

And

I seriously despise the method of CPU testing these days. Because you cannot see the difference properly. Honestly, i think they do this to create a market for both processors. Because when the FX series goes toe-to-toe with the i5's on low settings, it's shameful just how bad they are.

This is the part that I don't understand. I get that you want Intel to beat the living crap out of things, but the FX x3xx series isn't bad. And this proves that they are good enough, because they are.

I really do not see the problem here. They show that CPU does not matter in that game. Case closed.

[Xuvial]

Nooo keep it open, keep the case open! What have you done Tetris

[DeltrusDisc]

Hey.
Hey-hey-hey.

You can't just go and post that here. It's wildly off-topic. You get right back here and review that beer!
You can't mention Oktobeardfest and leave out the beer. There was beer, right? Else I'm thoroughly disappointed.

(else, I have no issues with the post)

My utmost apologies!

I personally only managed to get two beers that night, though I am DEFINITELY returning this coming Thursday for a special Dogfish Head release party for the 120 Minute IPA! I had to depend on my bud this night for beer, seeing how I couldn't get my check deposited from work until tomorrow, yes, I'm pissed off about that, and I am broke otherwise. He graciously bought me a couple, seeing as I drove us the whole 40 miles there and 40 miles back! Plus, had I not had a friend invite me to go, he would not have known, either!

I started my night out with a lovely snifter of Bell's Brewery's 2012 Vintage Expedition Russian Imperial Stout. One sniff and taste of that and I just proclaimed, "holy shit!" Chocolate, coffee, all the robust malty flavors with alcohol burns to finish, that a Russian imperial stout should have. It was truly a one of a kind beer! Absolutely DELICIOUS!

The other beer, which I would say was far less impressive but still tasty, was a crazy-named Allagash I could not begin to figure out, but it was essentially a rather fruity saison-style, by the flavor and mouth-feel of it.

Cheers!



(Kamelot is a band from Tampa, Florida, with singer from Sweden! )

- - - Updated - - -

Oh, I suppose I should add, this was the reward for the best in show!



The President or one of the guys behind Chicago Beardsmen brews his own beer and basically said, hey, this could be a fancy reward! And so it was.

[Majesticii]

You can't force your opinion and then close the case. Essentially saying that whatever i'm going to say next is irrevant since you're right, and i'm wrong. I didn't get the memo saying that being a forumadmin makes your opinion in arguments any more esteemed. So please, no 'holier than thou' act.

Well, they kind of do, in the regard that there is no performance difference under a scenario such as the one in the test. "Relative performance"? In that scenario, none. Reaching 60 FPS in a newer game? No, not even some of the best hardware out there today will do that. In say, Star Citizen.

If star citizen relies on IPC, a haswell will yield about double the framerate in regards to a FX x3xx. As does it in RTS/MMO games, which you agreed on. But you expect people to know this? They don't mention that anywhere in the review. They only say that Civ5 is an exception, but blame it on poor CF/SLI scaling. Not that it's an RTS, which means heavy CPU load is causing the poor scaling.

And This is the part that I don't understand. I get that you want Intel to beat the living crap out of things, but the FX x3xx series isn't bad. And this proves that they are good enough, because they are.

I never said they were bad (ok maybe i did, but hear me out), they're 'sufficient'. But they're not even remotely the same as ivy/haswell i5's. Which this review is trying to suggest. That is my problem. And then people on this forum start complaining when they get shitty FPS in WoW. Which makes me sad, because they spent their hard earned cash in a system that wasn't suited for their task. Caused by misinformation provided by the current benchmarking system which says "cpu doesn't matter".

And it is especially annoying when you consider the pricedifference between the FX-8350 and the i5-4670K. It is virtually non-existant. So why would you opt for the 8core having 50% less IPC.

- - - Updated - - -

Unrelated, ASUS denied my service because the card doesn't originate from my country, but from a neighbouring country (Germany). Probably the most lousy customer support i've ever witnessed.

[Faithh]

Looking at Majesticii's Civ 5 benchmark it was kinda obvious that Intel allowed to feed an extra gpu while there was totally no point to have SLI/CF for an AMD system. So basically if a random guy who mainly plays RTS/MMO games added a 2nd card 580 for civ 5 he would literally have wasted a lot of money and worst of all he can't take advantage of it. So where can we fit "AMD better for its price" when we had a price difference of 10-20$?

I'm hearing a lot of reasons that AMD does exceptionally well or even better in BF3/BF4 but I don't believe this at all, mainly because "BF3 is heavily multithreaded". Yeah right, my 3930K never passes 35% orsomething in BF3 and being a major bottleneck to my 780 making me to play on medium or low settings to try to have atleast 120fps all the time.

The reasons you hear nowadays why they bought an AMD is totally ridiculous especially if Intel has more headroom for better cards in the future in games that are very light threaded which we still are going to see. It wouldn't be really nice that you buy a new gpu and you get the same frames just because your CPU bottlenecked it.

[Majesticii]

Looking at Majesticii's Civ 5 benchmark it was kinda obvious that Intel allowed to feed an extra gpu while there was totally no point to have SLI/CF for an AMD system. So basically if a random guy who mainly plays RTS/MMO games added a 2nd card 580 for civ 5 he would literally have wasted a lot of money and worst of all he can't take advantage of it. So where can we fit "AMD better for its price" when we had a price difference of 10-20$?

I'm hearing a lot of reasons that AMD does exceptionally well or even better in BF3/BF4 but I don't believe this at all, mainly because "BF3 is heavily multithreaded". Yeah right, my 3930K never passes 35% orsomething in BF3 and being a major bottleneck to my 780 making me to play on medium or low settings to try to have atleast 120fps all the time.

The reasons you hear nowadays why they bought an AMD is totally ridiculous especially if Intel has more headroom for better cards in the future in games that are very light threaded which we still are going to see. It wouldn't be really nice that you buy a new gpu and you get the same frames just because your CPU bottlenecked it.

I mostly agree, but one thing is wierd though. Dropping graphical quality when your CPU is being the bottleneck shouldn't really effect the framerate at all. Unless you dropped things like physics and shadows.