Gaming Performance: E2180 vs. E4500

[Elfear]

I'm trying to pick a cpu for a build I'm doing. I've read a couple reviews that show a stock E2160 about 5-15% behind a stock E4300 in various games. For that big of a difference I'll probably fork over the extra $40 for the E4xxx series chip. I plan on getting 3.2-3.4Ghz out of whatever I buy. Question is, will an oced E2180 still be cpu limited at 1680x1050 vs. an oced E4500?

 

[SerpentRoyal]

Don't think the extra cache would amount to 10% improvement in games. Bottleneck is GPU if CPU is north of 3.0GHz.

 

[JackMDS]

Read this page, http://www.anandtech.com/cpuch...owdoc.aspx?i=3112&p=14

Or better off, read the whole Article.

Even better, recommend to every one to take a look at it.

 

[zach0624]

The 100% increase in cache will give you a 10%(200-300mhz) boost if the game is cpu limited, but at 1680x1050 games don't get cpu limited.

 

[SickBeast]

Originally posted by: JackMDS
Read this page, http://www.anandtech.com/cpuch...owdoc.aspx?i=3112&p=14

Or better off, read the whole Article.

Even better, recommend to every one to take a look at it.

I found that article misleading. They rag on the 2XXX series chips, when really if you extrapolate a 2140 chip at 3.2ghz, it would have been the fastest CPU they tested in ALL of the games (just double the 1.6ghz scores).

I'm not sure how scientific that is, but really they seem to take more of an issue with the 1MB of cache than they should.

 

[ArchAngel777]

IRRC, the Conroe was ~25% clock for clock faster than AMD. Allendale was around 15% and the Pentium E series is only about 5% clock for clock. Cache makes a huge different for the Core 2's... Strip them of this vital resource and AMD has a competing product.

IIRC, Xbit labs did some extensive testing and found that the E2X series overclocked to 3.33Ghz was still slightly outperformed by the C2X6800 (2.93Ghz).

Anyway, CPU limited and bottlenecking isn't an on or off switch like some believe.

For starters, I used to play at a lower resolution of 1280 x 768 with a A64 @ 2.5Ghz on a 8800GTS 640 OC2 (620/2000) and wasn't bottnecked in anything created within the last 2 years... Start moving back to 3Dmark 01 or games of that genre and even the fastest of CPU's are bottlenecked.

Quite honestly, I recommend reading at Xbit labs and legionhardware and just read through their reviews and be analytical about it.

 

[bryanW1995]

let's just assume that pentium E is even clock for clock with x2. if you take a 2180 and oc it to 3.3, it will still be lightning fast. take off 20% for the missing cache and that still gets you e6750 stock performance, which will dominate any game out there at much higher res than 1680x1050 if you have a decent video card.

 

[Elfear]

Here is one of the reviews I was talking about: Techspot. At 1600x1200 and 8xAA there is still a significant difference between the different C2D chips. Techspot only tests at stock clocks though so I don't know how much of a difference there would be at the 3.2-3.4GHz level.

Here is an excerpt from the review:
"This means that the ideal budget processor for gamers would be the Core 2 Duo E6320 at around $165, regardless of whether or not you plan to overclock.

Those that cannot afford a $165 processor should look at the E4300, which costs just $114 and was faster than the E2160 in all our gaming tests. While a 10% performance margin might not seem like much with a GeForce 8800 GTX, it is far more noticeable with a less powerful card like the GeForce 8600 GTS, for example. Therefore saving $20 on a processor is not really a smart idea for gamers. "

 

[21stHermit]

Originally posted by: ElfearHere is an excerpt from the review:
"This means that the ideal budget processor for gamers would be the Core 2 Duo E6320 at around $165, regardless of whether or not you plan to overclock.

Since I'm curious by nature. Can anyone explain why games are so sensitive to the amount of cache?

Most applications (encoding, compiling, photo processing), GHz rules, but for games its all about cache and video.

Just curious,
Hermit

 

[SerpentRoyal]

Originally posted by: Elfear
Here is one of the reviews I was talking about: Techspot. At 1600x1200 and 8xAA there is still a significant difference between the different C2D chips. Techspot only tests at stock clocks though so I don't know how much of a difference there would be at the 3.2-3.4GHz level.

Here is an excerpt from the review:
"This means that the ideal budget processor for gamers would be the Core 2 Duo E6320 at around $165, regardless of whether or not you plan to overclock.

Those that cannot afford a $165 processor should look at the E4300, which costs just $114 and was faster than the E2160 in all our gaming tests. While a 10% performance margin might not seem like much with a GeForce 8800 GTX, it is far more noticeable with a less powerful card like the GeForce 8600 GTS, for example. Therefore saving $20 on a processor is not really a smart idea for gamers. "

Hmmm...if you believe that report, then E4300 pulls 75.2, while E2160 72.3. That's only 4% difference! You need to learn how to interpret data.

At default 1.8GHz core speed, the E2160 is nothing special. Take this baby to +3.0GHz and the onboard cache becomes much less significant. Core speed is KING! A fast CPU with tons of cache cannot overcome a lousy GPU. If you need an extra 3 to 4 FPS to play a game, then buy a better GPU.

 

[JackMDS]

Originally posted by: SerpentRoyal

Hmmm...if you believe that report, then E4300 pulls 75.2, while E2160 72.3. That's only 4% difference! You need to learn how to interpret data.

You mean you actually have to look at the number?

You can not make a "wishful" decision if one line is a little shorter than the other?

 

[Madwand1]

Originally posted by: Elfear
Here is one of the reviews I was talking about: Techspot. At 1600x1200 and 8xAA there is still a significant difference between the different C2D chips. Techspot only tests at stock clocks though so I don't know how much of a difference there would be at the 3.2-3.4GHz level.

Here is an excerpt from the review:
"This means that the ideal budget processor for gamers would be the Core 2 Duo E6320 at around $165, regardless of whether or not you plan to overclock.

Those that cannot afford a $165 processor should look at the E4300, which costs just $114 and was faster than the E2160 in all our gaming tests. While a 10% performance margin might not seem like much with a GeForce 8800 GTX, it is far more noticeable with a less powerful card like the GeForce 8600 GTS, for example. Therefore saving $20 on a processor is not really a smart idea for gamers. "

I thought it was a pretty biased editorial comment there, essentially trying to make the point that you should give more money to Intel for their higher-priced CPUs.

Let's see: One argument they make is: If you buy a cheap CPU, you're probably going to buy a cheap motherboard, because it wouldn't make sense to spend more money on a motherboard than a CPU, and cheap motherboards suck. So there.

Now, obviously if you were to purchase an E2140 or E2160 processor it would be best to pair it with an ultra-cheap motherboard that supports it. There is little point buying a mid-range motherboard which will probably cost slightly more than the processor. Unfortunately most cheap motherboards offer very little flexibility when it comes to upgrading and also offer very few features.

Uh, maybe you'll be able to afford a better motherboard if you don't blow the money on a more expensive CPU?..

The other claim, that CPU is much more noticeable with a lesser card like an 8600 GTS isn't substantiated here, and is counter-intuitive. If you have a lesser graphics, the odds are greater that you'll be GPU-bound, not CPU-bound. Besides, if you're hitting 20 fps, is 10%- 22 fps really a big deal? And how can they claim that the 10% they saw with a high-end GPU scales down to low-end GPUs with the same processors without substantiation?

 

[PliotronX]

Look at it like this, the E4500 is about 50% more expensive. Since you will be overclocking to the same clockspeed, will the E4500 deliver 50% more performance?

Talking games here, 1680x1050 is not going to show a lick of a difference in cache sizes. Zach0624 said as much several posts ago. If you want the icing on the cake, go for it.

 

[bryanW1995]

Originally posted by: JackMDS

Originally posted by: SerpentRoyal

Hmmm...if you believe that report, then E4300 pulls 75.2, while E2160 72.3. That's only 4% difference! You need to learn how to interpret data.

You mean you actually have to look at the number?

You can not make a "wishful" decision if one line is a little shorter than the other?

I like graphs. They are so pretty.

 

[zach0624]

Originally posted by: PliotronX
Look at it like this, the E4500 is about 50% more expensive. Since you will be overclocking to the same clockspeed, will the E4500 deliver 50% more performance?

Talking games here, 1680x1050 is not going to show a lick of a difference in cache sizes. Zach0624 said as much several posts ago. If you want the icing on the cake, go for it.

I are smart!

This doesn't mean getting the e4500 is not warranted. It will get you a higher garunteed OC which is why I got it. I figured even if I got a bad chip it would reach 300mhz fsb. The only thing that really matters with these cpus is that they are at or above the game's recommended spec. For awhile I used a XP 2600+ but it had started to slow me down so I upgraded to a athlon 64 x2 and a core 2 duo.