An Intel processor through the eyes of an AMD user

[Tech Tweaker]

Admittedly, it's been a long time since I ran an Intel rig in my own home and had the ability to test it at will, like as in 5-6 years. The last Intel I owned was a Pentium III 866MHz. For the past five years or so my main (and only) desktop rig has been my trusty AMD socket 754 single-core with a 1.8GHz Sempron 3000+. So, you can probably imagine how me venturing into the Intel section of the CPU forum is kind of the same as visiting a foreign country.

Today I went about testing some socket 775 hardware I had here, just because I wanted to see if it all would play nicely together given it was all obtained in used condition and some of it was untested when I obtained it.

First I discovered that apparently HWMonitor can't read the right temps on a C2D chip, which annoyed me a bit since that's what I normally use, and so then I had to go download RealTemp.

After I got that out of the way though, I discovered that this thing really does run cool; okay, that's a lie, I actually discovered that before I downloaded RealTemp, because I was putting my hand on the heatsink with HWMonitor going and saying to myself "there's no way it's that hot." Such was why I decided to download this RealTemp thing, since I've overheard people from the Intel camp talk about it in HW reviews in various places, to see if the temps the two report would differ, and they differ greatly.

The second thing I discovered was that these Intel chips apparently cool down really FAST. I literally had this thing at full load and it peaked at 50°C, less than two minutes later I had the heatsink off and after putting the back of my hand to both the heatsink and the processor I discovered they both weren't at all warm to the touch, if anything they were room temp or less. Now, the AMD socket 939 rig actually has its load temps about the same, but it seems to take longer to return to the normal idle temp range.

You might be asking yourself why I removed the heatsink that quickly. That would be because I neglected to put any paste on the heatsink when I first installed it, and proceeded to forget that fact. Plus when I saw HWMonitor showing temps of 60°C on both cores I was concerned I was about to fry the processor, little did I know at the time though that it was 10°C higher than the actual temps. A secondary reason was that I was wondering just how hot one of these would get without any paste on the heatsink, and how much cooler it would run with some paste applied.

Another thing which surprised me was the speed at which this chip was able to complete a run of SuperPi 1m. Now, my AMD socket 939 rig with a dual-core A64 4200+ 2.2GHz chip running at all stock speeds can complete a run in around 39-40s, this C2D E4400 (2GHz) socket 775 chip completed it in 28s. In all honesty I expected my 939 rig to beat the 775 by at least a little bit, given that it does have a 200MHz advantage in clock speed, or at the very least to give the 775 a run for its money. That didn't happen here though, as the result was quite the opposite with the socket 775 completely trouncing the 939 in nearly half the amount of time.

These tests were performed with both machines having stock cooling in the CPUs, and running 2GB of value ram each, with the 939 running with a SATA HDD and the 775 using an IDE (it was what I had handy, and it already had a stripped down OS loaded onto it, good enough for testing purposes anyway).

Now, I'm impressed by what I've seen today, but I'm still not sure it justifies Intel's price tags.

 

[MattNo5ss]

You might be asking yourself why I removed the heatsink that quickly. That would be because I neglected to put any paste on the heatsink when I first installed it, and proceeded to forget that fact. Plus when I saw HWMonitor showing temps of 60°C on both cores I was concerned I was about to fry the processor, little did I know at the time though that it was 10°C higher than the actual temps. A secondary reason was that I was wondering just how hot one of these would get without any paste on the heatsink, and how much cooler it would run with some paste applied.

You have to remember that Intel chips run hotter than AMD, and can handle hotter temps. Intel chips usually don't reach thermal throttling temps until ~90-100C depending on the CPU. So as long as you're below throttling, then you should be fine. Typical safe temps for 24/7 are ~75C in Prime95. However, I've run my i5 680 up to 105-110C with thermal throttling disabled.

Another thing which surprised me was the speed at which this chip was able to complete a run of SuperPi 1m. Now, my AMD socket 939 rig with a dual-core A64 4200+ 2.2GHz chip running at all stock speeds can complete a run in around 39-40s, this C2D E4400 (2GHz) socket 775 chip completed it in 28s. In all honesty I expected my 939 rig to beat the 775 by at least a little bit, given that it does have a 200MHz advantage in clock speed, or at the very least to give the 775 a run for its money. That didn't happen here though, as the result was quite the opposite with the socket 775 completely trouncing the 939 in nearly half the amount of time.

Yeah...AMD and SuperPi???

Intel chips beat AMD chips in all 2D benchmarks really.

Now, I'm impressed by what I've seen today, but I'm still not sure it justifies Intel's price tags.

AMD = performance per dollar, Intel = strictly performance. So, Intel can charge whatever they want to since they don't have competition in the performance category. Hopefully new AMD chips will help drive Intel's prices down. I really want a good CPU for 3D benchmarks, but those CPUs are so expensive...

 

[Bisquit]

I know why I went from Amd to intel, the noise. Always had noisy boards with amd cpu's, hsf's made noise also. Heat, always a lot of heat, running around 40° idle, fiddling with all kinds of case fans to lower that.

I must say, intel has the faster chips since they went core2 and up. My gaming experience likes that. Hsf's for intel are pretty quiet, overclocking goes better/easier, less heat at idle etc.
Still, we got 2 phenom rigs in the house, they run pretty smooth. With noise ofcourse.

 

[RollingThunder]

Admittedly, it's been a long time since I ran an Intel rig in my own home and had the ability to test it at will, like as in 5-6 years. The last Intel I owned was a Pentium III 866MHz. For the past five years or so my main (and only) desktop rig has been my trusty AMD socket 754 single-core with a 1.8GHz Sempron 3000+. So, you can probably imagine how me venturing into the Intel section of the CPU forum is kind of the same as visiting a foreign country.

Today I went about testing some socket 775 hardware I had here, just because I wanted to see if it all would play nicely together given it was all obtained in used condition and some of it was untested when I obtained it.

First I discovered that apparently HWMonitor can't read the right temps on a C2D chip, which annoyed me a bit since that's what I normally use, and so then I had to go download RealTemp.

After I got that out of the way though, I discovered that this thing really does run cool; okay, that's a lie, I actually discovered that before I downloaded RealTemp, because I was putting my hand on the heatsink with HWMonitor going and saying to myself "there's no way it's that hot." Such was why I decided to download this RealTemp thing, since I've overheard people from the Intel camp talk about it in HW reviews in various places, to see if the temps the two report would differ, and they differ greatly.

The second thing I discovered was that these Intel chips apparently cool down really FAST. I literally had this thing at full load and it peaked at 50°C, less than two minutes later I had the heatsink off and after putting the back of my hand to both the heatsink and the processor I discovered they both weren't at all warm to the touch, if anything they were room temp or less. Now, the AMD socket 939 rig actually has its load temps about the same, but it seems to take longer to return to the normal idle temp range.

You might be asking yourself why I removed the heatsink that quickly. That would be because I neglected to put any paste on the heatsink when I first installed it, and proceeded to forget that fact. Plus when I saw HWMonitor showing temps of 60°C on both cores I was concerned I was about to fry the processor, little did I know at the time though that it was 10°C higher than the actual temps. A secondary reason was that I was wondering just how hot one of these would get without any paste on the heatsink, and how much cooler it would run with some paste applied.

Another thing which surprised me was the speed at which this chip was able to complete a run of SuperPi 1m. Now, my AMD socket 939 rig with a dual-core A64 4200+ 2.2GHz chip running at all stock speeds can complete a run in around 39-40s, this C2D E4400 (2GHz) socket 775 chip completed it in 28s. In all honesty I expected my 939 rig to beat the 775 by at least a little bit, given that it does have a 200MHz advantage in clock speed, or at the very least to give the 775 a run for its money. That didn't happen here though, as the result was quite the opposite with the socket 775 completely trouncing the 939 in nearly half the amount of time.

These tests were performed with both machines having stock cooling in the CPUs, and running 2GB of value ram each, with the 939 running with a SATA HDD and the 775 using an IDE (it was what I had handy, and it already had a stripped down OS loaded onto it, good enough for testing purposes anyway).

Now, I'm impressed by what I've seen today, but I'm still not sure it justifies Intel's price tags.

Tech,

Both AMD and Intel are fine products, I've owned both equally over the years.

Overall Intel pricees aside, I went from an AMD Opteron 180 to this Intel E8500, both overclocked to the max on stock cooling to within safe limits.

32MB SuperPi Opteron 180: ~ 29 mins.
Intel E8500: ~ 12m 30 secs.

I understood the Intel would be faster but I didn't realize it would be that much faster. The Opteron served me well and I never had any complaints about it.

Intel temp sensors leave much to be desired at idle but both of these average out under load. CPU operating temps for the Intel are higher but I was prepared for that beforehand.

 

[Tech Tweaker]

You have to remember that Intel chips run hotter than AMD, and can handle hotter temps. Intel chips usually don't reach thermal throttling temps until ~90-100C depending on the CPU. So as long as you're below throttling, then you should be fine. Typical safe temps for 24/7 are ~75C in Prime95. However, I've run my i5 680 up to 105-110C with thermal throttling disabled.

AMD = performance per dollar, Intel = strictly performance. So, Intel can charge whatever they want to since they don't have competition in the performance category. Hopefully new AMD chips will help drive Intel's prices down. I really want a good CPU for 3D benchmarks, but those CPUs are so expensive...

I wonder then why I've always been told that AMD's ran hotter than Intel's.

Maybe everyone that had been telling me that was referring to how they were in the old days.

I know I've seen some of these reviews of heatsinks or paste where they talk about Intel chips hitting the low to mid 70's°C, and to me that just seems insanely hot; being that I normally only own AMD processors that is all I have for a point of reference to compare temps to and at those temps the average AMD would have to essentially be bursting into flames (not literally).

I wonder why the Intel chips run so hot...

 

[Theocnoob]

I wonder then why I've always been told that AMD's ran hotter than Intel's.

Maybe everyone that had been telling me that was referring to how they were in the old days.

I know I've seen some of these reviews of heatsinks or paste where they talk about Intel chips hitting the low to mid 70's°C, and to me that just seems insanely hot; being that I normally only own AMD processors that is all I have for a point of reference to compare temps to and at those temps the average AMD would have to essentially be bursting into flames (not literally).

I wonder why the Intel chips run so hot...

No idea. I know you can take the core temperature of a bloomfield i7 to near 100 degrees without killing it. I've personally seen over 80 degrees on a 4Ghz i7. It doesn't do any harm. Just a different design. There's a smaller temperature delta too. An AMD chip that idles 30 will only hit about 45 loaded while an Intel chip idling around 25 might hit 60 loaded. Doesn't matter, performance wise, at the end of the day.

 

[EarthDog]

Definately a green (as in little exposure) point of view!

 

[Theocnoob]

Definately a green (as in little exposure) point of view!

That the heat output isn't necessarily an indication of or related to compute power?

 

[EarthDog]

Im not sure what heat has to do with price/performance.

But if I did a quit Froogle product search, looks like that A64 3800+ (s939) costs nearly 10% more as the e4400.

Not sure what both prices were at their launch though.

Mattno hit it right on the head.

I wonder then why I've always been told that AMD's ran hotter than Intel's.

Maybe everyone that had been telling me that was referring to how they were in the old days.

Who knows, but you were told incorrectly.

S939 AMD beat out Pentium4 and P4D chips on s775. When C2D came out, AMD lost the clock for clock performance crown. Maybe thats what they were comparing it against...Old s775 stuff?

 

[muddocktor]

Back in the ancient past of the Pentium III and AXP, that held true about Thunderbird AMD running hotter than Intel. But at the same time, the Thunderbird AXP also had a slightly higher IPC than the Pentium III and also overclocked better generally. It wasn't until the Tualatin P III came out that Intel could hang with AMD on overclocking. And I'm not gonna even talk about that abortion called the Willamette P 4, which was pretty much crap.

 

[Bisquit]

"When C2D came out, AMD lost the clock for clock performance crown."

So true, this was for me the moment I jumped on the bandwagon called intel.
In my experience it was the idle temps of amd chips that always were a little bit higher compared to intel cpu's, but then, my last intel at that point was a pIII. Full load was usually in the low 60's when overclocking, mid 50's at default speed with my amd chips.
It was at a time when amd cpu's were faster because of their internal clock that ticked three times faster or something like that. Anyway, loved my amd64 xp3800+.
Price when I bought it was about the same as for i5 2500k now, roughly.

I've seen my 2600k hit 98°C. Times sure have changed with that, although I'm still caught in the high 50's low 60's movement for full load temperatures.

 

[theELVISCERATOR]

Amd is cheap, Intel is fast, take your pick.

 

[stunt]

Funny I should come across this thread.

I'm on my Pentium 940D "legacy" machine right now. With the Thermalright 120mm HP cooler and a gentle typhoon 1850 it idles at 43c and spikes up the 65c at a blink of an eye in a Lian-Li G70B full tower case. I can't seem to let this system go because I like to go back and see how things were in 2003. With a 945pvs mobo and 3 gigs of 667 ram along with 2x 80g 7200s in raid 0, it runs Win7x64 very well. ( I have a HD6850 in it for HDMI use).

I have flogged this system from day one (8 years now) and it won't die. Just keeps chugging along. I think I got my moneys worth.

 

[boucher91]

aw heck i still have a p pro200@233 been running as such for bout 10+yrs..

 

[freeagent]

Im pretty sure the heat comes from the cache. Just like in the old days

I had some sweet a64 setups back before conroe.. I didnt believe conroe was as strong and as fast as they said, but sure enough after I killed yet another nvidia (abits and asus)board, I had enough and thought Id try out an intel.. havent looked back since

 

[ChanceCoats123]

For all we know, the temps might be exactly the same on the die. However, placement of the sensor(s) as well accuracy are very important. Also, I can tune my Fluke 51 to read -196*C in a 0*C environment. Doesn't mean it's true. For all we know, Amd might use a -40*C offset so they read lower max temps or maybe intel uses a +40*C offset so the max temps run a bit higher. IMO, it doesn't much matter with ln2.

 

[Tech Tweaker]

I guess in retrospect I probably should have expected the Intel setup to be faster, after all, Intel processors are quad-pumped (four instructions per clock-cycle), whereas AMD's are double-pumped (two instructions per clock-cycle). I'm not sure why I didn't think of that at the time, I knew of this, but it merely slipped my mind.

I guess the higher level of throughput going through the processor might explain the extra heat, come to think of it.

 

[EarthDog]

Im sure the 'pumps' have something to do with it, but wouldnt expect better performance just on that premise alone...

Again with throughput, that may have something to do with heat, but what about the P4 vs Athon64 when A64 thumped them, the P4 was a notorious heater and it was slower...

 

[Tech Tweaker]

Im sure the 'pumps' have something to do with it, but wouldnt expect better performance just on that premise alone...

Again with throughput, that may have something to do with heat, but what about the P4 vs Athon64 when A64 thumped them, the P4 was a notorious heater and it was slower...

Well, the Intel S775 setup was also running dual channel PC2-4200 memory, whereas the AMD S939 setup was running dual channel PC3200 memory.

Now that one I think I can explain.

At that time the P4 required a much higher voltage to operate (up to 1.75V on some models, I know because I owned one), thereby resulting in high levels of heat, whereas the A64's were requiring 1.35-1.4V.

 

[EarthDog]

So to that theory, fast forward through generations and see if it holds true (it doesnt).

AMD's are what, ~1.35v now? While Sandybridge for example is ~1.16? C2D duals are usually ~1.25 as well.

Remember voltage vs clockspeed in regards to temperatures, voltages have a TREMENDOUSLY LARGER effect on temps that clock speed or instructions /cycle.