Intel Core 2 Quad Q9300 Processor - HotHardware

Intel Core 2 Quad Q9300 Processor

2 thumbs up

While a 95W TDP rating doesn’t seem extremely power-friendly, this TDP rating is actually quite good in comparison to other quad-core solutions on the market. AMD’s new Phenom 9850 chips run in the range of 125W TDP, and Intel’s previous generation quad-cores run art 105W+. While the spec sheets will tell you one story, we can say that after testing this chip first hand, it’s very friendly in terms of power consumption and heat-production for a quad-core. The Q9300 doesn’t need large, copper contraptions with fancy heatpipes and high-speed fans to stay cool (although, they help). The chip is perfectly happy running on fairly basic cooling systems with low-to-moderate speed fans. When idling, our Q9300 chip ran just above room temperature, and only hit temperatures of roughly 125°F / 51°C under sustained processing loads. For a quad-core with this much computing power under the hood, that's pretty impressive.

Of course, this makes for a beautiful segue to talk about overclocking. As the Q9300 runs cool under sustained loads at its stock 2.5 GHz clock rate, it’s only natural to think that the chip can run at much higher clock speeds when pushed. Being the low-end model of a new family of processors, people will be looking at this chip to be the new low-cost, high-speed overclocking price/performance champ that the Q6600 once was, and still is. From our tests thus far, the Q9300 does indeed surpass the overclockability of the Q6600, but not by the huge amounts we had originally expected. Here’s why.

The Q9300 runs with a 7.5x multiplier at 1333 MHz front side bus, which isn’t ideal for overclocking. The multiplier can be adjusted downwards, but it’s not incredibly flexible, leaving potential overclockers to rely on high front side bus speeds. Most modern motherboards don’t have any trouble hitting 1666 MHz FSB speeds and up, but when you get to the areas of 1800-1900 MHz and up, you really have to rely on the motherboard in order to help you unlock the true potential of this chip. With our tests, with a brand, spanking new XFX nForce 790i SLI motherboard, we were able to coax a maximum clock speed of 3.4 GHz out of our (2.5 GHz stock-clocked) Q9300 sample processor. Early reports from users getting these chips in hand are showing similar results.

In order to reach clock speeds of 3.3 GHz and higher, the Q9300 requires a voltage boost (up to 1.325V) in order to stay happy. Even with high clock speeds and increased voltage levels, our chip ran surprisingly cool throughout our tests. At 3.4 GHz, our chip would run through all of our benchmarks but one (a video encoding benchmark), no matter what voltage or cooling we threw at it. We found the sweet spot of overclocking on the Q9300 to be 3.33 GHz, which was rock solid throughout all of our testing. We’ve included a set of benchmarks at this overclocked speed in the following pages.


Core 2 Quad Q9300 - Stock Speeds


Core 2 Quad Q9300 - 3.33 GHz O/C

One of the most interesting areas to look at is the Core 2 Quad Q9300’s power consumption. With Intel’s 45nm process technology, a fairly low clock speed, and a small amount of cache, we would expect power consumption levels to be somewhat low in comparison to other quad-core solutions on the market today. We ran our power consumption numbers with our standard testbed (seen in the following page) with identical hardware configurations. Power consumption levels are tested for the whole system (not just the processors) through a hardware AC wattage meter. Idle numbers are taken at the Windows desktop after the OS has loaded and settled down. Full load meters are taken during our Cinebench R10 test when all four cores are active and running at maximum consumption levels. Power management features (EIST, C1E, Speedstep, PowerNow) were disabled across the board in order to keep consistency.


At first glance, the Core 2 Quad Q9300 looks very good from a power consumption perspective.  This 45nm quad-core chip at 2.5 GHz only consumes a scant few more watts compared to Intel’s 45nm dual-core E8500 at 3.16 GHz.   The Q9300 consumes about 30W less under load compared to the Q6600, despite a higher clock speed, which is also impressive.

What is most impressive is that even when overclocked to 3.3 GHz, the power consumption of this processor is roughly at the same level as a Core 2 Quad Q6600 processor at a stock 2.4 GHz.  Even with nearly a 1 GHz clock speed difference there is nearly no difference in power consumption, which is a testament to Intel's Penryn microarchitecture and its advanced manufacturing prowess.

Article Index:

1 2 Next
0
+ -

It looks to be a good value overclocked, but the Q6600 is so much cheaper I still have to recommend it for people looking for a good entrance into Intel Quad Core. 

0
+ -

Looks good, but yeah I would opt for the Q450.. ;)

0
+ -

The now lower price of the Q6700 makes it a very viable option for the enthusiast/gamer with it's 10 multi... personally I'd go with that if I were in the market for a quad right this minute.

Q6700 G0 stepping FTW!!!

0
+ -

The statement that this one is worthy upgrade from Q6600 making this review to look like commercial promotion.

Sorry guys but this chip can't give you 3.4-3.6 like the Q6600.

The strange 7.5 multiplier is kind of limitation for overcloking. 

The only good thing is ~30 watt power saving that will save me ~30$ a year.

Actually this Q9300 is about the same price as Q6700 which has higher stock clock. It is interesting why it was not in test. Direct performance for $ compare would be more informative about advantages of the newbie.

 If you already have quad, stay with it and wait for real improvements with next very soon coming architecture with integrated memory controller and whatever new socket.

It can be recomended for people that must buy new computrer for some reason, but with almost NO REAL IMPROVEMENTS over older slower  Q6600, you just can't be serious  recommending it as upgrade.

 Technical note:

 On the third page (Test System Details) , it is stated that  you used  eVGA Nvidia nForce 680i LT SLI Motherboard (For Intel Testing).

Is there any changes and 680i started to support 45nm quads ? 

0
+ -

Hi there :

As for an upgrade vs. the Q6600, we wouldn't really recommend people taking this route.   However, if one was buying a new system and the Q9300 vs. Q6600 was up in the air, the Q9300 is likely a better shot in this time.  Better raw performance, lower power consumption, SSE4 instrucuctions.  Both chips are highly overclockable - the Q6600 has some advantages, but the Q9300 can certainly overclock quite a bit too.

The eVGA nForce 680i LT motherboard does indeed support 45nm dual and quad-core processors with the latest BIOS.  We used the same platform for this article as we did for the 45nm "Wolfdale" E8500 article, which works great. 

0
+ -

“Power management features (EIST, C1E, Speedstep, PowerNow) were disabled across the board in order to keep consistency.” By consistent you mean that that the power figures are consistently unrepresentative of what real world shipping PCs will achieve! I’d much rather see figures with all the power saving features enabled and if you want to add a second graph with them turned off then fine. One advantage of that is that you see more clearly how much more your system consumes at idle when you over-clock by increasing VCore in the BIOS which leaves your VCore permanently set at the higher value. The difference can be dramatic which is why I only over-clock using the stock VCore which means EIST is still fully functioning and idle power draw is much lower. It typically limits me to 3GHz with an Intel Quad but it seems a good compromise to me.

0
+ -

agharta - Thanks for the comments - I'll look into adding power consumption numbers w/ power conservation modes enabled for future articles.

 

0
+ -

If I were to choose between a Q6600 and a Q9450 or a Q9550 I would choose for sure a Q9450 or 550.

If I were to choose between a Q6600 and a Q9300 I would surely choose the Q6600 because of its low price point and very little difference in performance.Q9300 features a lower quantity of cache than it's 45nm colleagues.

0
+ -

Ok.
There are a lot of reports that  680i  not supporting  45nm  qauds.

But luckily (almost no extra cost) i laid my hands on Q9300 yesterday. My mobo is Asus P5N32-E SLI (stripped down Striker. Same PCB, same components, only few things missing like LCD, Buttons and little bit different chipset cooling).

In the beginning, i just replaced Q6600 and pressed the power button. I got an error in POST, that the CPU is unrecognized and suggestion that i should update my 1302 BIOS. Quick check of BIOS and found funny things. The CPU multiplier was integers only. Cache size was reported as 1x2K.  but everything else seemed to be ok. I booted to OS (Gentoo Linux 64 bit) and did some tests (mostly compilations and compression) it was working fine. The strange thing was the temperature. It was reported in BIOS as 31C and so in OS, but the cores temp was close to 60C. I tried Vista 64 bit and it was fine. I just wanted to update BIOS from GUI but it failed, so i copied the 1404 BIOS to flash drive and used BIOS utility to update it form the BIOS. After update, system just refused to boot. No beeps, no sign of life only HDD fans spinning. I tried to remove memory, no change. So (being very upset) i placed back Q6600 and it did the boot. I put back the Q9300 and it worked. CPU was correctly identified by BIOS, the CPU multiplier has now only 3 options: 6, 7 and 7.5. Cache reported correctly. I experienced some weird issues with BIOS settings, but you cant expect everything to work with beta version. But after the update, CPU temperature reported in BIOS as 60C !!! I think it's incorrect because the cooler remained pretty cool near the base. I use scythe ninja for long time and if the processor getting hot, you can verify it by heatpipes temperature near the base. So I started to play a little with clock. Results with default voltages are: FSB - 1650MHz, CPU - 3.09Ghz (2.5 stock), Mem - 825MHz (800 stock) @ 4-4-3-15 1T (linked sync mode). 

That was the first day testing, i'm doing migration from 250GB mirror of sensitive data to 750GB mirror. I left it for the night, and it was completed by the morning with very impressive 52MB/sec. 

The questions are:

1. What should I do  with temperatures ? can BIOS report it incorrectly ?

2. What else should I check in order to verify that the processor is compatible with my mobo ?

0
+ -

with the temp. i would get a maybe Infrared Thermometers to get the temps for yourselfs cause i heard that a couple of nvidia mobo were having trouble with geting temps, from the new 45nm duals and quads

1 2 Next
Login or Register to Comment
Post a Comment
Username:   Password: