Intel Pentium4 3.06GHz with Hyper-Threading 14th November 2002

Today we have great pleasure in being amongst the first to bring you a review of Intel's latest CPU offering - the Pentium4 3.06GHz with Hyper-Threading. Not only is it the fastest Pentium available but it is the first to bring Jackson technology to the masses. If you want the lowdown on Hyper-Threading you can find our comprehensive guide here. We've had a complete system from Intel for about a week now and have been busy benchmarking for our readers' pleasure. Here's some CPU info:

Note that the CPU is not fully recognized as it is a pre-production sample. As with nearly all images in our reviews you can click on them to see a bigger version so don't worry if something looks unreadable - clicking on it will make things clearer. On to the benchmarks.

Test Setup

Three systems will be used for our testing purposes:

Sisoft Sandra

Let's get the synthetic benchmarks out of the way first. The tests with Hyper-Threading enabled are on the left and on the right is the same test without Hyper-Threading.

We can see improvements in the CPU benchmarks and as expected no improvement in memory bandwidth (for obvious reasons). These improvements are not to be sniffed at and are roughly equivalent to a couple of speed grades.

PC Mark 2002

Again the Hyper-Threaded benchmark is on the left and the one without on the right.

Surprisingly this benchmark shows a performance decrease in one area and a slight increase in another. Don't worry, this is a problem with the way the benchmark is written and it will be updated soon to support Hyper-Threading. This benchmark does not appear to make use of multi-threading so cannot benefit from the extra logical processor.

3D Mark 2001

As usual, the Hyper-Threading benchmark is on the left.

The settings used are the default ones. This is 1024x768, 32-bit color, compressed textures, D3D Pure Hardware T&L, no antialiasing. This shows that this benchmark does not benefit from Hyper-Threading. Does this means that games will not benefit? Let's take a look at Unreal Tournament 2003.

Unreal Tournament 2003

Let's take a look at the flyby and botmatch figures at various resolutions.

And here are the results from the botmatch benchmarks.

The figures are very impressive in that they show this is the fastest CPU money can buy but they also show that Hyper-Threading is not really going to help much in games until they take advantage of Dual Processors. Game designers would like to see more Dual Processors before they program for them so we have a chicken and egg situation. Things may have changed now that Hyper-Threading will be a feature on all Intel Pentium4s from now on in that software companies will no longer be able to consider Dual CPU users as a minority since in effect any chip with Hyper-Threading will have two logical processors.

Divx Encoding

Let's turn to an area where we know that Hyper-Threading and Dual CPUs will make a difference - Audio and Video encoding. This is becoming more and more popular and is very computationally intensive with long processing times (relatively speaking that is, this field is not for those that complain about how long their Outlook Express takes to load).

For consistency we will use Jet Li's The One as our test matter. It is not interlaced and contains a mixture of action types and is not too long. There will be three tests all using Divx 5.02. Audio will be encoded separately. I will try and keep my commentary to a minimum as all configuration information is shown in the images below.

Xmpeg 4.5

Firstly we will start with the industry standard Xmpeg with its settings at those recommended by Intel and as used by other sites such as Toms Hardware. Here are the CODEC settings:

and here are the Xmpeg settings:

and here are the results

There is a gain of over 23% from Hyper-Threading, thoroughly trouncing the Dual Athlon machine I was until now so impressed with. Both CPUs are made good use of as you can see by the single CPU machine's result in comparison.

AviSynth and VirtualDub

No serious Divx encoder uses Xmpeg alone and it's just used by the media for benchmarking purposes so let's get serious. We ripped our source material to hard disk and created a DVD2AVI project file using forced film (it was 99% film). Loading this into Gordian Knot we first saved an .avs file with no changes at all (720x480) and no filters of any sort. This was loaded into VirtualDub with the following CODEC parameters :

After encoding we got these results:

Even without Hyper-Threading the P4 system was beating our Dual Athlon rig! Benefits from Hyper-Threading are 30% on the first pass, 27% on the second and 28% on average.

This is all good stuff but how about a real-world test? To simulate a realistic test we added a neutral bicubic resize filter in the .avs file  and used the following CODEC parameters (including two popular Pro settings) which are designed to total 700MB (when the audio is muxed in):

Which resulted in the following.

Here we see Hyper-Threading bringing an overall increase in performance of some 15%, comfortably ahead of the Dual Athlon system. This increase is not as great as in the previous test but represents a more realistic encoding process from the point of most Divx users, and the CODEC only benchmarks can almost be regarded as synthetic (this is not actually true as they are practical benchmarks but it's an interesting comparison).

Audio Encoding

What about audio? We took the AC3 track from the above sample material and used HeadAC3he to convert it into Vorbis format so our final muxed file could have Ogg containment. There isn't space here to go into the advantages of Ogg Vorbis over MP3 and AVI so let's just say that Vorbis sounds about the same as MP3 for half the file size or twice as good for the same file size (that is subjective though).

These are the settings we used:

Since it is more meaningful to show throughput than time taken (which depends on the length of the source) we display the results thus:

We're not really using both CPUs efficiently here but despite this the Hyper-Threading of the P4 has managed to claw an extra 20% bonus over what it would be otherwise.

Conclusion

The Pentium4 just keeps getting better and better and with Hyper-Threading we are getting a free taste of what it's like to be the owner of a Dual CPU system. It will take some time for performance in games to appear but there is a significant boost where time really matters and now we have a viable single CPU alternative to that "cheap Dual Athlon system" that used to compare so favorably price-wise to a high-end Pentium4. Why buy a Dual Athlon when a single Pentium4 can give you better results? The ball is firmly in AMDs court and they seem to be falling further and further behind. I am assured that Intel will not be resting on their laurels despite this, and they will continue on schedule with their aggressive roadmaps.

With the upcoming release of Dual Channel DDR chipsets from Intel we may be in for a new Golden Era in home computing. If you want to know how Hyper-Threading works and exactly what you need to make sure your system takes full advantage of Hyper-Threading you can find our Guide to Hyper-Threading Technology right here.

All trademarks are the property of their respective owners.