FinalWire has launched a major update to its AIDA64 benchmarking and diagnostic utility, bringing entirely rewritten memory benchmarks and support for AMD and Intel's latest processor products.
AIDA64 3.00, released today, brings with it entirely new memory benchmarks which replace previous versions - described by the company as 'outdated.' 'The new bandwidth benchmarks now use multiple threads to squeeze out every last bit of performance from the caches and the memory modules. On modern multi-core processors, using the old single-threaded benchmarks you couldn't see the actual memory bandwidth, but only the memory bandwidth available for single-threaded applications,' the company explains of its new release.
'With the new benchmarks you will however get considerably higher scores, much closer to the theoretical memory bandwidth available. It is especially true for 3-channel and 4-channel memory configurations, such as Intel X58 and X79 based high-end desktop systems; and also for NUMA-enabled multi-socket systems, such as 2- and 4-way AMD Opteron and Intel Xeon based servers and workstations.'
Figures quoted by the company show a dramatic difference: an Intel Core i7-3960X running on an X79 chipset motherboard with quad-channel DDR3-1600 memory jumps from 16,825MB/s memory read using the old-style benchmark code found in AIDA64 2.85 to a whopping 45,640MB/s in AIDA64 3.00. Similar boosts can be found in cache memory testing, with multi-threading and AVX/AVX2 support seeing figures tripled compared to the old method. While this provides a more accurate estimation of the capabilities of modern multi-threaded memory subsystems, it does mean that scores are in no way comparable between the new AIDA64 release and previous versions.
The new software also includes rewritten memory latency benchmarks, using a new block-random approach that prevents memory controllers from over-optimising the results in order to avoid reading too-low figures. Again, this means that results won't be comparable between versions - although those lucky enough to have their hands on Intel's highest-end Crystal Well-equipped Haswell processors with their embedded L4 eDRAM cache will be pleased to hear that there's support for benchmarking that part of the system, too.
Further optimisations include support for processor and chipset information dumping for Intel's Haswell and AMD's Kabini and Temash families and updated support for the latest Nvidia and AMD graphics hardware.
As usual, FinalWire is offering a 30-day trial of AIDA64 Extreme Edition, the home-user oriented release, from its official website. Those who want to use it beyond that time will need to pick up a licence at £28.13, while commercial users will need an Engineer Edition licence with 50 per cent discounts being offered to educational establishments and students.
AIDA64 3.00, released today, brings with it entirely new memory benchmarks which replace previous versions - described by the company as 'outdated.' 'The new bandwidth benchmarks now use multiple threads to squeeze out every last bit of performance from the caches and the memory modules. On modern multi-core processors, using the old single-threaded benchmarks you couldn't see the actual memory bandwidth, but only the memory bandwidth available for single-threaded applications,' the company explains of its new release.
'With the new benchmarks you will however get considerably higher scores, much closer to the theoretical memory bandwidth available. It is especially true for 3-channel and 4-channel memory configurations, such as Intel X58 and X79 based high-end desktop systems; and also for NUMA-enabled multi-socket systems, such as 2- and 4-way AMD Opteron and Intel Xeon based servers and workstations.'
Figures quoted by the company show a dramatic difference: an Intel Core i7-3960X running on an X79 chipset motherboard with quad-channel DDR3-1600 memory jumps from 16,825MB/s memory read using the old-style benchmark code found in AIDA64 2.85 to a whopping 45,640MB/s in AIDA64 3.00. Similar boosts can be found in cache memory testing, with multi-threading and AVX/AVX2 support seeing figures tripled compared to the old method. While this provides a more accurate estimation of the capabilities of modern multi-threaded memory subsystems, it does mean that scores are in no way comparable between the new AIDA64 release and previous versions.
The new software also includes rewritten memory latency benchmarks, using a new block-random approach that prevents memory controllers from over-optimising the results in order to avoid reading too-low figures. Again, this means that results won't be comparable between versions - although those lucky enough to have their hands on Intel's highest-end Crystal Well-equipped Haswell processors with their embedded L4 eDRAM cache will be pleased to hear that there's support for benchmarking that part of the system, too.
Further optimisations include support for processor and chipset information dumping for Intel's Haswell and AMD's Kabini and Temash families and updated support for the latest Nvidia and AMD graphics hardware.
As usual, FinalWire is offering a 30-day trial of AIDA64 Extreme Edition, the home-user oriented release, from its official website. Those who want to use it beyond that time will need to pick up a licence at £28.13, while commercial users will need an Engineer Edition licence with 50 per cent discounts being offered to educational establishments and students.
MSI MPG Velox 100R Chassis Review
October 14 2021 | 15:04
Want to comment? Please log in.