Alexey Shobanov

"Finally, the fact that we were waiting for so long!" It is such or similar in the sense of exclamations here is the third month of computer publications, devoting more and more new reviews and testing core processors 2 Duo, about whose exit company Intel announced in mid-July. The emergence of these chips based on the new microarchitecture Intel Core, without exaggeration, it became the main event of the current year, which demonstrated to all that the "genre crisis" was overcome and all the problems associated with the previous microchitectural machine Intel Netburst are behind. Naturally, I did not remain aside and our magazine - on his pages a number of articles telling both the features of the architecture of new processors and chipsets and motherboards designed to work with them were published. But, unfortunately, until recently, we have failed to spend a direct comparison of new processors from Intel with solutions from their main competitor - Advanced Micro Devices (AMD). Moreover, AMD, anticipating Core 2 Duo, a little less than two months before the announcement of the new Intel architecture (May 23, 2006) introduced its new AM2 platform (as described in detail in detail on the pages of our magazine). Its basis was the processors built on already well and long-known microarchitecture AMD64, but in this case Having a built-in memory controller capable of maintaining a DDR2 SDRAM memory and made in a new form factor with AM2 processor connector. Today, finally, we had the opportunity to reduce competitors face to face and evaluate their capabilities when performing a wide range of tasks. For comparison, we have chosen Intel Core 2 DUO E6600 processors and AMD Athlon. 64 x2 5000+, and that is why: both models have about the same price. So, the AMD Athlon 64 x2 5000+ processor in the batches from a thousand pieces cost $ 301, and Intel E6600 - $ 316. In addition, today both chips take the same place in model row Companies producers, being the second model in the appropriate rules. In tab. 1 shows some key characteristics of these chips.

Table 1. Characteristics of AMD Athlon 64 x2 5000+ and Intel Core 2 DUO E6600

CPU	AMD Athlon 64 x2 Dual-Core	Intel Core 2 Duo
Frequency, MHC
System tire, frequency, MHz / Bottom, GB / C	Hypertransport / 2000/8	Quad Pumped Bus / 1067 / 8.5
Number of Cores
Maximum temperature, ° С
L1 Cache instructions, KB
L1 Cache data, KB
L2 Cache, Krib
Energy Saving Technology	Cool'n'quiet.	Enhanced Intel Speed \u200b\u200bStep
SIMD instruction set
TechProcess
Processor connector

In order to compare the possibilities of these two processors, we took advantage of a number of specialized programs, as well as test scenes and scripts for popular applications, which allowed us to evaluate performance computer Systemsbased on these central processors when performing various tasks. Here are a list of used tests and applications:

• pC performance as a whole:

Crystal Mark 9.0;

• scientific calculations: Science Mark 2.0;
• audio encoding: LAME 3.98A;
• video encoding:

XMPEG 5.2 Beta + DivX Converter 6.2.5,

Windows Media Encoder 9,

TMPGENC 2.524,

MainConcept MPEG Encoder 1.51,

MainConcept H.264 Encoder v.2.0;

• office applications:

VERITEST BUSINESS WINSTONE 2004 V.1.0.1,

VERITEST BUSINESS WINSTONE 2004 MULTITASKING TEST V.1.0.1,

VERITEST MULTIMEDIA CONTENT CREATION 2004 V.1.0.1;

• archiving:

• gaming tests:

Doom 3 (Path 1.3),

Far Cry (Patch 1.33),

Quake 4 (Patch 1.05, SMP-Enable);

• working with 3D graphics:

Discreet 3DS MAX 7.0 (Specapc 3DS MAX 7 v.2.1.3 Script),

Alias \u200b\u200bWavefront Maya 6.5 (Specapc MAYA 6.5 V1.0 script);

Digital photo processing: Adobe Photoshop. CS2.

For testing two stands were collected:

• for AMD Athlon 64 x2 5000+ processor:

Motherboard - ASUS M2N32-SLI Deluxe (chipset - NVIDIA NFORCE 590 SLI),

Hard Drive: Seagate Barracuda 7200.7 with a capacity of 120 GB (ST3120827AS), NTFS file structure;

• for Intel Core 2 DUO E6600 processor:

Motherboard - ASUS P5B Deluxe (Chipset - Intel P965 Express),

Video card - Sapphire Radeon 1900 XTX,

System memory - 2xcorsair CM2X512-8500 in DDR2-800 SDRAM mode (total 1 GB), timings 4-4-4-12 (CAS Latency-Ras to Cas Delay-Row Precharge-Active to Precharge),

Hard disk - Seagate Barracuda 7200.7 with a capacity of 120 GB (ST3120827AS), NTFS file structure.

Testing was conducted under operating system Microsoft Windows. XP SP2 with ATI Catalyst 6.7 video driver installed.

Let us turn to the results obtained by us during testing (Table 2). According to the results of the synthetic tests of Futuremark PCMark 2005 and CrystalMark 9.0, allowing to evaluate the work of individual subsystems of the computer system, we see that the performance of the processor subsystem and the configuration memory subsystem is based on intel processor Core 2 Duo E6600, 10-15% higher than that of similar subsystems based on AMD Athlon 64 x2 5000+. At the same time, podestes relating to other subsystems (disk and graphic) did not reveal any significant advantage of the Intel platform, except for the OPENGL test OGL Crystalmark 9.0, in which, however, the geometry with intensive loading is calculated. central processorso say that it is in pure form The test of the graphic subsystem is impossible. Moreover, in two other tests on the schedule of the same test package - GDI and D2D - the platform on AMD Athlon 64 x2 5000+ had a tangible advantage compared to a competing solution. This situation has developed in podtes that assess the performance of the disk subsystem: According to the results of the HDD Futuremark PCMark 2005 test, for the computer system based on the AMD processor, it was the same for both computer systems, and the results of the HDD Crystalmark 9.0 test turned out to be 12% higher than for the platform Intel. From all this, you can make a very important conclusion: when you hold all subsequent tests, the obtained difference in performance between compared configurations (if we are talking about the fact that the advantage is on the side of the Intel platform) is determined primarily by the capabilities of the processor subsystem and a bundle "Processor - Memory" Since neither the graphic nor the disk subsystem in this case has no advantage over the competing solution.

Table 2. Results of testing AMD AMD Athlon 64 x2 5000+ and Intel Core 2 DUO E6600

			AMD Athlon 64 x2 5000+	Intel Core 2 Duo E6600	Difference (%)
Price, dollars
Futuremark PCMark 2005.
Science Mark 2.0.	Molecular Dynamics.
	Memory Benchmarks.
Audio encoding (LAME 3.98A), with
Video encoding
	Windows Media Encoder 9 (AVI -\u003e WMV), with
	TMPEGENC 2.524 (AVI -\u003e M2V + WAV), with
	MainConcept H.264 Encoder v.2.0 (AVI -\u003e MPG), with
	MainConcept MPEG Encoder v.1.51 (AVI -\u003e MPG), with
VERITEST BUSINESS WINSTONE 2004 V.1.0.1
VERITEST BUSINESS WINSTONE 2004 V.1.0.1 Multitasking Test
VERITEST MULTIMEDIA CONENT CREATION WINSTONE 2004 V.1.0.1
Archiving	7-zip 4.42 (size size 64 MB, word length 256 KB), with
	WinRar 3.51 (Compression Method: Normal), C
	HDR / SM 3.0 Score
Half-Life 2, resolution 1024x768
Doom 3 (Path 1.3), resolution 1024x768
Far Cry (Patch 1.33), resolution 1024x768
QUAKE 4 (Patch 1.05, SMP-Enable), resolution 1024x768
Discreet 3DS MAX 7.0 + Specapc 3DSMAX7 V.2.1.3 (Software Render)
ALIAS WAVEFRONT MAYA 6.5	(Specapc Maya 6.5 v1.0)
Adobe Photoshop CS2, with

In the queue, the SCIENCE MARK 2.0 utility test set, designed to estimate the performance of the system when performing scientific computing. By contacting the results of these tests, it is not difficult to note that when performing scientific calculations (Podtes of Molecular Dynamics, Primordia and Cryptography) The advantage of AMD Athlon 64 x2 5000+ looks very convincing. This result is quite explained, since it has long been known that operations with a floating point (on which all the calculations performed in this case are based) are the skate of AMD processors with the K8 generation core, as, however, and K7. Although in this case it is very curious that with a pure synthetic test on the execution of floating-point operations BLAS / FLOPS (calculating special matrices size from 64x64 to 1536x1536) The processor from Intel turns out to be faster!

Another set of tests, where the AMD Athlon 64 x2 5000+ processor managed to take the top over the Intel processor, became the VERITEST 2004 package, which simulates the user with office applications (VERITEST BUSINESS WINSTONE 2004 V.1.0.1), as well as the creation of Internet content (Veritest Multimedia Content Creation Winstone 2004 v.1.0.1). It can be assumed that in this case a small advantage of the AMD platform is caused by a little better work Disk subsystem and greater processor clock frequency (2.6 versus 2.4 GHz at Intel Core 2 Duo E6600). At the same time in the multisascy test when working with office applications (VERITEST Business Winstone 2004 v.1.0.1 Multitasking Test) The Intel platform turns out to be more productive. Most likely, one of the reasons for this has become more efficient use of second-level cache (L2), which is general, and not individual for each of the cores (as it is implemented in the processors of the AMD Athlon 64 x2 family) and also has four times more Volume (4 MBs against the total 1 MB of AMD ATHLON 64 x2 5000+).

On the coding tasks of video and audio files and archiving, the Intel Core 2 Duo E6600 processor has proven to be much faster than the platform on AMD Athlon 64 x2 5000+ - its gain here was from 4.4 (MainConcept H.264 Encoder v.2.0) to 24, 5% (MainConcept MPEG Encoder v.1.51). Moreover, this advantage is achieved by the Intel processor no longer at the expense of a higher clock frequency, as it was with the NetBurst architecture processors, and due to the best organization of working with streaming data ...

And what about games? Until recently advantages when testing gaming applications unconditionally remained for aMD processors. And it was on his field that the Advanced Micro Devices brainchild was for this time a crushing defeat. In all game tests, the first platform was the platform on the Intel Core 2 Duo E6600, and the advantage was very significant on test scenes (from 21% in stage for Quake 4 to 38.8% for Half-Life 2).

In tests that assess the performance of the system when working in popular 3D-packages Discreet 3Ds Max 7.0 and Alias \u200b\u200bWavefront Maya 6.5, as in the execution of the script that imitating the work of the user by processing digital photo In Adobe Photoshop CS2, the advantage of the Intel Core 2 Duo E6600 processor over the competitor also does not cause any doubt.

Thus, according to the results of this comparison, it is possible to state: new Duo 2 Core processors of Intel, built on the basis of the Intel Core microarchitecture, today are significantly superior to the level of performance of competitors, the only serious of which is Advanced Micro Devices. In addition, it can be said that the enemy in the face of AMD Athlon 64 x2 5000+, who spoke in the role of the continuer of the glorious business of processors with AMD64 microarchitecture, was a bit of its weapon. So, abandoning the race at frequencies, the embodiment of which the processors of the family became Intel Pentium. 4 With their microarchitecture NetBurst, Intel made a bet in Intel Core to increase the number of operations performed on the batter and optimize computing. It is also interesting that the Intel Core 2 DUO E6600 bypasses its opponent not only in pure performance, but also in all its relative expressions: relative performance per unit cost and relative performance per unit of power. Recall that the TDP processor Intel Core 2 Duo E6600 is 65 W, and the level of power dissipation in AMD Athlon 64 x2 5000+ is 89 W. Directly compare these values, of course, not quite correctly, because to determine their company use different techniques, however, for some approximate comparison, they are quite suitable.

Product release date.

Lithography

Lithography Indicates semiconductor technology used for the production of integrated chip sets and the report is shown in a nanometer (NM), which indicates the size of the functions built into the semiconductor.

Number of Cores

Number of nuclei is the term hardwaredescribing the number of independent central processing modules in one computing component (crystal).

Number of streams

A stream or stream of execution is a software term denoting the basic orderly sequence of instructions that can be transmitted or treated with one CPU core.

Basic processor clock frequency

The basic processor frequency is the speed of opening / closing processor transistors. The base frequency of the processor is the operating point where the calculated power (TDP) is set. Frequency is measured in gigaherts (GHz) or billions of computing cycles per second.

Cache memory

Cool memory processor is a high-speed memory area located in the processor. Intel® Smart Cache's Intelligent Cache Memory Indicates an architecture that allows all the kernels to jointly dynamically use access to the lature cache.

System Tire Frequency

The tire is a subsystem transmitting data between computer components or between computers. As an example, you can call a system bus (FSB), which is exchanged between the processor and the memory controller block; DMI interface, which is a point-to-point connection between the built-in controller intel memory and Intel I / O controllers block system board; And the Quick Path Interconnect interface (QPI) connecting the processor and the integrated memory controller.

The parity of the system tire

The parity of the system tire provides the ability to verify errors in the data sent to the FSB (System Bus).

Calculated power

Calculated thermal power (TDP) indicates the average performance in watts when the processor power is dissipated (when working with a base frequency, when all kernels are involved) under complex load, a specific Intel. Familiarize yourself with the requirements for thermoregulation systems presented in the technical description.

SCENARIO DESIGN POWER (SDP)

Max. Calculate Power is an additional reference point of thermoregulation, designed to use devices associated with high temperature, with the simulation of real operating conditions. It balances requirements for performance and power during workloads throughout the system, and provides the most powerful use of the world's use. Talk to technical Description Products for full information on capacity specifications.

Voltage range VID

The VID voltage range is the indicator of the minimum and maximum voltage values \u200b\u200bon which the processor should work. The processor provides VID with VRM (Voltage Regulator Module), which, in turn, provides the correct voltage level for the processor.

Available options for embedded systems

Available options for embedded systems indicate products that provide extended acquisition for intelligent systems and embedded solutions. Product Specification and Terms of Use are presented in the PRODUCTION REPEASE Qualification (PRQ) report. Contact an Intel to get detailed information.

Supported connectors

The connector is called a component that provides mechanical and electrical connections between the processor and the motherboard.

T Case.

Critical temperature is the maximum temperature allowed in the processor integrated heat distributor (IHS).

Intel® TURBO BOOST ‡ Technology

Intel technology® Turbo Boost dynamically increases the processor frequency to the required level using the difference between the nominal and maximum values Parameters of temperature and energy consumption, which makes it possible to increase the efficiency of energy consumption or, if necessary, "dispersed" processor.

INTEL® HYPER-THRAADING ‡ Technology

Intel® Hyper-Threading Technology (Intel® HT Technology) provides two processing streams for each physical core. Multi-threaded applications can perform more tasks in parallel, which significantly speeds up performance.

Intel® Virtualization Technology (VT-X) ‡

Intel® Virtualization technology for directional I / O (VT-X) allows one hardware platform to function as a few "virtual" platforms. The technology improves management capabilities, reducing downtime and maintaining productivity through the allocation of individual sections for computing operations.

Intel® 64 ‡ Architecture

Intel® 64 architecture in combination with the corresponding software Supports the operation of 64-bit applications on servers, workstations, desktops and laptops. The Intel® 64 architecture provides an increase in productivity, due to which computing systems can use more than 4 GB of virtual and physical memory.

Set of commands

A set of commands contains basic commands and instructions that the microprocessor understands and can perform. The value shown indicates which set of Intel commands this processor is compatible.

Prohibition status

The idle status mode (or C-status) is used for energy saving when the processor is inactive. C0 means working condition, that is, the CPU in this moment Performs useful work. C1 is the first state of inaction, C2 - the second state of inaction, etc. The higher the numerical indicator of the C-state, the more the program performs the program on energy saving.

Enhanced Intel Speedstep® Technology (Improved Intel SpeedStep® Technology)

Intel SpeedStep® advanced technology allows you to provide high performance, as well as compliance with the requirements mobile systems to energy saving. Standard Intel Speedstep® technology allows you to switch the voltage and frequency level depending on the processor load. Intel SpeedStep® enhanced technology is built on the same architecture and uses such development strategies such as separation of voltage and frequency changes, as well as the distribution and restoration of the clock signal.

Intel® Demand Based Switching Technology

Intel® Demand Based Switching is a power management technology, in which the application voltage and the clock frequency of the microprocessor are held at the minimum required level until the computing power is required. This technology was presented on the server market called Intel Speedstep®.

Technology thermocontrol

Thermocontrol technologies protect the processor housing and a system from a failure due to overheating using multiple control functions temperature regime. The intracrystalline digital temperature sensor (Digital Thermal Sensor - DTS) determines the kernel temperature, and the temperature control function can reduce the power consumption of the processor, thereby reducing the temperature, to ensure work within normal performance.

New Intel® AES Commands

Intel® AES-NI commands (Intel® Aes New Instructions) are a set of commands that allows you to quickly and safely provide encryption and data decoding. The AES-NI commands can be used to solve a wide range of cryptographic tasks, for example, in applications providing group encryption, decoding, authentication, generation of random numbers and authenticated encryption.

Bit Performance is a hardware security function that reduces vulnerability to viruses and malicious CodaAnd also prevent malware and distribute it on the server or online.

After the release of the previous material about the new Intel processors, it took not so much time, so this article It will be more logical to perceive not as an independent, but as a kind of supplement. It so happened that the Intel Core 2 Duo E6600 processor fell into our hands after the first article out. Of course, in itself, he is not very interesting because It differs from Core 2 Duo E6700 only by a multiplication coefficient (and, respectively, by 266 MHz less frequency). Of course, it would be much more interesting to test the E6300 / 6400 with the "Upolyn" cache, or even the youngest E4200 in the lineup, which also cut the tire to 800 MHz. Unfortunately, these CPUs have not yet reached us. Therefore, for the absence of the most desirable, we suggest you while reading another material on the topic "The performance of the new Intel architecture in widespread real tasks." Fortunately, the topic could not be bored - this is just the second material dedicated to it :). Hardware and Software

Configuration of test posted stands

CPU.	Mainboard.	Memory.
Athlon 64 FX-62	(BIOS 9.03)	Corsair CM2X1024-6400 (5-5-5-12)
Athlon 64 FX-60	EPOX EP-9NPA3 (BIOS 06.03.30)	Corsair CMX1024-3500LLPRO (2-3-2-6)