New Intel dual-core beats AMD quad

EddieC 0 534 Views

Intel tomorrow will officially unveil a pair of dual-core mobile processors with its Hyper threading technology, adding to its x86 line a low cost chip with performance akin to that of the company's i7, Pentium 4 and Xeon parts. The Core i3 and Core i5 processors, demonstrated by Hewlett-Packard and others at through Sunday, also incorporate graphics functions on board, saving power and cost by eliminating the need for a separate graphics package.

In a benchmark report published yesterday, the two-core i5 661 processor outperformed a quad-core Athlon II 640 running at 3GHz.With on-board graphics turned off, the i5 can overclock "well beyond 4GHz," according to the report. A six-core version is expected in the next few months. Formerly code-named Clarkdale for desktops and Arrandale for laptops, the Core i3 and Core i5 are the first processors available using Intel's 32nm process.

Only the i5, which is equipped with a feature called Turbo Boost, can automatically overclock to 2.26GHz based on processing and power efficiency requirements. The integrated GPU still uses a portly 45nm die, and will only work when used in a system that's also equipped with Intel's new H55 or H57 chips. The new chips also include a DDR3 memory controller. Loaded laptops built around the a Core i5 chip are expected to initially sell for around US$900, while i3-based machines will go for about $700.

Dani AI

Generated

A couple of quick points that help make sense of 's report and stop one benchmark from becoming a buying decision.

Benchmarks favor different things. Many games and common apps still spend most time on a single or a few threads, so a CPU with stronger per-core performance (higher clock or better instructions-per-cycle) can beat a chip with more physical cores in those tests. Conversely, heavily parallel tasks (media encoding, rendering, large compiles, server workloads) scale with core count and will usually favor the quad‑core. Cache size, memory bandwidth, and platform features (chipset, PCIe lanes, firmware) also change results more than sheer core number in many scenarios.

How to read or reproduce a claim like this:

  • Prefer real-world tests that match your workload (x264/HandBrake for encoding, Blender/Cinebench for rendering, a CPU-limited game test for gaming).
  • Use identical system components (same GPU, RAM config, BIOS defaults) and update chipset/GPU drivers before testing.
  • Run each test 3+ times and record averages; monitor CPU utilization and temperatures to see throttling.
  • Note whether background services or OS power plans are influencing results; isolate the test machine as much as possible.

Practical buying guidance: match the platform to the primary use case. If you do heavy multitasking or multi-threaded work, prioritize more physical cores and platform expandability. If you mostly game or run single-threaded apps, single-thread speed and a good GPU matter more. Finally, pay attention to cooling, BIOS updates, and driver support — inconsistent firmware or thermal throttling can flip real-world results regardless of advertised specs.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.

Edit Preview