An Introduction Of Sisoftware Sandra

Published Date: 02 Nov 2017

SiSoftware Sandra, stands for the System ANalyser, Diagnostic and Reporting Assistant, is a system information and diagnostic software which gives us detailed information about the software that stored on our systems and the hardware that our computers are made with. Sandra provides undocumented information about software and hardware as well as other devices whether they are software or hardware by analyzing, diagnosing and benchmarking computers’ components.

The program offers a variety of benchmarks to test the processor speed, speed of our hard drives, file system, removable storage, network performance, video rendering, mobile device file transfer, memory bandwidth and a whole lot more. All result values are especially useful when we want to optimize our systems with few justifications of hardware or software.

Sandra enables user to draw both a high and low levels comparisons by providing more comprehensive system information, such as the CPU, chipset, video adapter, ports, printers, sound card, memory, network, Windows internals, AGP, PCI, PCIe, ODBC Connections, USB2, 1394/Firewire, etc.

With this system diagnostic tool, it helps enhancing our theoretical learning from textbook and lecture notes by providing us with a plenty of detailed system information either software or hardware. The diagnosed data provides us with a better understanding of the computer architecture.

Part 2: Analysis on Sandra Report

Sandra Report allows us to compare products of the same product type (such as PC, laptop, smart phone, etc.) which is either from different companies/manufacturers or products are made with different components. Sandra software provides users a collection of standardized measurements on hardware, software and benchmark. Based on the data generated, we can make comparisons among the devices, and give a conclusion on which part of device is better over the other one. Sandra Report plays an important role when we want to know more about our system or we want to improve our system performance.

Part 3: Definition for Rating Performance Terms

SPEC Benchmarks

The Standard Performance Evaluation Corporation (SPEC) is a nonprofitmaking organization whose aim is to establish and maintain a standardized set of benchmarks. These benchmarks can be used to analyze the latest generation of high-performance computers.

SPECCPU2006

SPECCPU2006 is designed to provide comparative computer-intensive performance measures across the wide range of computer systems. These measurements are all based on real user application’s workloads developed.

SPECint/SPECfp

SPECCint and SPECfp are computer benchmark measurements for measuring and comparing computer-intensive performance in terms of speed (how fast the workloads finished). SPECCint is used to examine the integer performance, whereas SPECfp is used to examine the floating point performance.

SPECintRate/SPECfpRate

SPECintRate and SPECfpRate are designed to examine and compare computer-intensive performance in terms of throughput ratio (number of tasks can be completed in parallel). SPECCint is used to study the integer performance, whereas SPECfp is used to study the floating point performance.

SPECJVM98/ SPECJBB

SPECJVM98 (Java Virtual Machine) is a benchmark for evaluating the Java Virtual Machines’ performance, and it can only compatible with JDK 1.1 API or later. When executing programs in a Java platform environment, it is suitable to use network and standalone Java client computers.

SPECJBB (Java Server Benchmark) is a benchmark for evaluating the Java server side performance. This measurement is done by emulating a three-tier client/server system with more focuses on the middle tier.

SPECPower

SPECPower is the first industry-standard benchmark for measuring the power/energy and performance features of volume server class computers. SPECPower indicator can help us analyzing power/energy efficiency which will provide us a right direction on how to minimize the power/energy consumption.

Dhrystone/ Whetstone

Dhrystone and Whetstone are synthetic benchmark for analyzing the performance of computers. Dhrystone, is a benchmark test, measures the speed and efficiency at which a computer performs integer operations, whereas Whetstone is performs floating-point operations.

OpenGL

OpenGL (Open Graphics Library) is the widely adopted 2D and 3D graphics API. It is concerned only with processing data in GPU memory, including rendering into a frame buffer and reading values stored in that frame buffer.

OpenCL

OpenCL is a benchmark which is designed for tasks running parallel on CPU, (GP) GPUs and hardware accelerators. The unit of OpenCL is pixels/s which means how many pixels can be computed in one second. The higher OpenCL value the better performance.

Efficiency Metrics

Performance vs. Cost (Cost Efficiency)

Cost Efficiency makes a comparison between the performance of the device and its purchasing cost, and it will provide us a ratio of performance (MOPS) over cost ($). The higher the ratio the better the cost efficiency, this is because we get the higher performance with the same amount of money spending.

Capacity vs. Power (Size Efficiency)

Size efficiency is a ratio of capacity (system’s memory (MB)) over power (energy consumption (W)). With this ratio, we are able to analyze how to fully utilize our system without wasting any of resources either memory capacity or power consumption. The higher the ratio of capacity over power the better the size efficiency.

Performance vs. Power (Power Efficiency)

Power Efficiency is a benchmark measurement that compares the performance of the device with its power. Power Efficiency is a ratio of performance (MOPS) over power (W), and the higher the ratio of performance over power the better the power efficiency.

Memory Bandwidth and Latency

Memory bandwidth refers to the size of the memory bus multiplied by the speed at which the memory is clocked. Higher memory bandwidth has faster data processing; this is because higher bandwidth has more memory buses.

Latency is a time delay in transmitting data. Latency is one of the most significant factors to affect CPU utilization. In order to increase system performance, reducing the waiting time is the most important issue to deal with rather than increasing the execution speed.

Cache/Memory Bandwidth

Cache/Memory Bandwidth is the rate at which data can be read from or stored into cache or memory by a processor. Cache is used to provide faster data transfer speed; this is because cache is nearer to CPU compare to the external memory. With the cache/memory bandwidth, we can analyze it and find out what is the possible reason that cause low CPU performance throughput.

Power Management Efficiency

A benchmark is used to analyze the efficiency of the power management technologies using by processors. More and more people are concerning of "Green" issue, hence this benchmark is useful in terms of analyzing how environmental friendly are the processors in the system. Lower power means lesser heat generated which means more environmental friendly.

Part 4: Evaluation on HP dm4-3090se System Performance

<<< Computer Overview >>>

< ID >

Host Name: YuHsin

Workgroup: WORKGROUP

< Computer >

Model: HP Pavilion dm4 Notebook PC 103C_5335KV G=N

L=CON B=HP S=PAV X=Null

Serial Number: 2CE21*****

Chassis: HP Notebook

Mainboard: HP 1813

Serial Number: PCNTD1A*******

BIOS: Insyde F.08 05/02/2012

BIOS (Basic Input Output System) is built-in software which determines what a computer can do without accessing programs from a disk. On personal computer, the BIOS contains all the code required to control the keyboard, display screen, disk drives, serial communications, and a number of miscellaneous functions.

Insyde Software company specializes in Unified Extensible Firmware Interface (UEFI) system firmware and engineering support services. InsydeH2O BIOS, is one of the company’s products, implements with Intel Platform Innovation Framework for UEFI/EFI.

Intel vPro: 8.00.02.1410

Total Memory: 7.9GB DDR3 SO-DIMM

< Processors >

Processor: Intel(R) Core(TM) i7-3612QM CPU @ 2.10GHz (4C 8T 2.8GHz/3.1GHz, 2.8GHz IMC, 4x 256kB L2, 6MB L3)

Processor is a computer’s brain, and it’s also known as Central Processing Unit (CPU). Processor deal with processing of numeric data, meaning information entered in binary form and the execution of instructions stored in memory.

Intel Core, is Intel company’s brand name, refers to microprocessors which is designed for mid-range to high-end customers and business.

The Intel Core i7-3612QM is Ivy Bridge architecture, and it is a fast quad-core processor for laptops based. It is the faster mobile processor rated at 35 Watt. The Hyperthreading allows its fore core to deal with up to eight threads in parallel; this will enhance a better utilization of the CPU.

Clock rate is number of clock cycles per second when a CPU is running. This laptop’s clock rate is 2.10 GHz; it means it can execute 2.1G cycles per second.

Socket/Slot: FC PGA988

< Chipset >

Memory Controller: HP Core (Ivy Bridge) Mobile DRAM Controller 100MHz, 2x 4GB DDR3 SO-DIMM 1.6GHz 128-bit, Integrated Graphics

Memory controller, is a digital circuit, controls the data flow from or to the main memory. Intel’s Core i7 integrates memory controller (IMC) on the microprocessor, this will help to reduce memory latency.

Ivy Bridge is one of the microarchitecture, and it is a codename line of Intel processors based on the 22nm manufacturing process. Ivy Bridge is an improved die shrink of the Sandy Bridge architecture with enhancements on both GPU and CPU performance.

DDR3 stands for Double Date Rate Type, and it is a SDRAM type which is used for system memory. DDR3 is offered in both DIMM and SO-DIMM form factors.

SO-DIMM (Small Outline Dual In-Line Memory Module) is about half length of a regular size DIMM. The small in size is suitable for laptop; this is because it has greater flexibility in designing the memory slots.

< Memory Module(s) >

Memory Module: Kingston 9905428-085A00LF 4GB DDR3 SO-DIMM

PC3-12800S DDR3-1600 (11-11-11-29 5-40-13-6)

Memory Module: Nanya NT4GC64B8HG0NS-DI 4GB DDR3 SO-DIMM PC3-

12800S DDR3-1600 (11-11-11-29 5-40-13-6)

< Video System >

Monitor/Panel: Samsung Generic PnP Monitor (1600x1200, 13.9")

Video Adapter: Intel(R) HD Graphics 4000 (16CU 16SP SM5.0 650MHz/1.1GHz, 1GB DDR3 1.6GHz/3GHz 128-bit, Integrated Graphics)

< Graphics Processor >

OpenCL GP Processor: Turks (480SP 6C 500MHz, 1GB)

Compute Shader Processor: Intel(R) HD Graphics 4000 (16SP 350MHz/1.1GHz,

1GB DDR3 1.6GHz/3GHz 128-bit, Integrated Graphics)

Compute shader is a programmable shader stage, and it provides high-speed general purpose computing. It gains benefits on the large number of parallel processors on the graphics processing unit (GPU). The computer shader offers thread synchronization features and memory sharing, this ensures to have more effective parallel programming methods.

<<< Processors >>>

< Processor >

Manufacturer: Intel

Model: Intel(R) Core(TM) i7-3612QM CPU @ 2.10GHz

Speed: 2.8GHz

Minimum/Maximum/Turbo Speed: 1.2GHz - 2.1GHz - 3.1GHz

Cores per Processor: 4 Unit(s)

Threads per Core: 2 Unit(s)

Type: Laptop/Netbook

Bus: Intel CSI (Quick Path)

Processor bus provides the path for communication between the CPU and the main bus. It is also a communication medium between the CPU and the processor support chipset which includes chips such as an external memory cache and the bus controller chip.

Types of processor buses:

Memory Bus transfers digital data and signal between the CPU and the RAM.

Address Bus transfers the next memory or I/O location address where data has to be delivered.

Input / Output Bus transfers input and output signals between CPU and other external peripheral devices.

Intel CSI (Common System Interface) is known as Intel QuickPath Interconnect, and it is a point-to-point processor interconnects. CSI refers to a width, point-to-point, packet-based interface implemented as two unidirectional links with low-voltage differential signaling.

Multiplier: 28x

Minimum/Maximum/Turbo Multiplier: 12x - 21x - 31x

Front Side Bus Speed: 100MHz

Front Side Bus (FSB) is a computer communication interface (bus) which connects the computer’s processor to the system memory (RAM) and other components on the motherboard. It is also called the "system bus"; this is because FSB serves as the main path from the processor to the rest of the motherboard. A CPU’s FSB speed is the speed at which data is processed in a computer. A slow FSB will result in the CPU to spend significant quantities of time waiting for data to arrive from system memory.

< Co-Processor (FPU) >

Speed: 2.8GHz

Type: Integrated

Revision/Stepping: 3A / 9

< Memory Controller >

Integrated in Processor: Yes

Memory controller integrates in processor will help reducing memory latency, hence resulted in increasing the system’s performance.

Speed: 2.8GHz

Multiplier: 28x

<<< Processor Arithmetic >>>

< Performance vs. Power >

Processor(s) Power: 35.000W

Aggregate Native Performance: 2517.26MOPS/W

Dhrystone Integer Native SSE4.: 3142.54MIPS/W

Whetstone Double Native SSE3: 2016.40MFLOPS/W

Results Interpretation: Higher Scores mean Better Performance.

In terms of power efficiency, the integer performance (3142.54MIPS/W) is more efficient than the floating point performance (2016.40 MIPS/W). This is because analyzing Dhrystone on a 64-bit processor with an integrated multiply-divide unit which will enhance the utilization of the compiler, thus improve the performance throughput.

< Performance vs. Speed >

Aggregate Native Performance: 28.48MOPS/MHz

Dhrystone Integer Native SSE4.: 35.56MIPS/MHz

Whetstone Double Native SSE3: 22.82MFLOPS/MHz

Results Interpretation: Higher Scores mean Better Performance.

For every MHz, Dhrystone integer performance also better than Whetstone double performance. Integer performance almost 1.5 times more efficiency than float-point performance.

< Processor >

Model: Intel(R) Core(TM) i7-3612QM CPU @ 2.10GHz

Speed: 3.09GHz (99%)

Type: Laptop/Netbook

L1D (1st Level) Data Cache: 4x 32kB, 8-Way, 64bytes Line Size, 2 Thread(s)

L1I (1st Level) Code Cache: 4x 32kB, 8-Way, 64bytes Line Size, 2 Thread(s)

L2 (2nd Level) Data/Unified Cache: 4x 256kB, ECC, 8-Way, 64bytes Line Size, 2 Thread(s)

L3 (3rd Level) Data/Unified Cache: 6MB, ECC, 12-Way, Fully Inclusive, 64bytes Line Size, 16 Thread(s)

Cache is designed to reduce the average waiting time for CPU to access main memory. Cache is used to store copies of most frequently used data from the main memory. There is a tradeoff between hit rate and cache latency; the larger caches have larger hit rates but longer latency. In order to balance of this tradeoff, many computers use multiple levels of cache from small fast cache (L1) to larger slower caches (L2 & L3).

L1 cache is a very fast memory built into the CPU to help improve computer performance. L2 and L3 cache are slower than L1 cache but faster than Random Access Memory (RAM).

Multi-level caches work by checking the smallest level 1 (L1) cache first. If it hits, the processor continues at high speed. If it misses, the next larger cache (L2) will be checked. The checking is done through level by level before external memory is checked.

8-Way is a fully pipelined architecture and permits consecutive hits and misses to be checked sequentially without stalling.

ECC (Error Detection and Correction) is a BIOS feature that performs detection and correction of single-bit errors in data stored in the L2 cache.

<<< Mainboard >>>

< Disk Controller >

Model: HP Mobile SATA RAID Controller

OEM Device Name: Intel Mobile SATA RAID Controller

Interface: SATA

SATA (Serial ATA) is a computer bus interface for connecting host bus adapters to mass storage devices such as hard disk drives and optical drives. SATA, is a serial link, a single cable with a minimum of four wires creates a point-to-point connection between devices. The thinner serial cables provides more efficient airflow inside a form factor and also allow for smaller chassis design.

Revision: A5

Maximum SATA Mode: G2 / SATA300

G2 / SATA300 refers to SATA revision 2.0 – 3 Gbit/s – 300 MB/s. G2 means second generation SATA interfaces which runs with a native transfer rate of 3.0 Gbit/s. SATA300 indicates that the maximum uncoded transfer rate is 300 MB/s (2.4 Gbit/s).

Mobile: Yes

Channels: 3

In Use Channels: 2, 67%

Port: G2 / SATA300

Port: G2 / SATA300

Part 5: Comparison on Two Systems’ Performance

HP Pavilion dm4

ASUS K43SJ

<<Processor>>

Model

Intel(R) Core(TM) i7-3612QM CPU @ 2.10GHz

Intel(R) Core(TM) i7-2630QM CPU @ 2.00GHz

Speed

3GHz (96%)

1.2GHz (41%)

Total Memory

7.9GB DDR3 SO-DIMM

8GB DDR3 SO-DIMM

Comments:

Performance Throughput

Both laptops are running on Intel Core i7 system. However, HP Pavilion dm4 is using i7-3612QM CPU which has higher clock speed (2.10 GHz) than ASUS K43SJ’s clock speed (2.00GHz). HP Pavilion dm4 will result in higher performance throughput than ASUS K43SJ. On the other hand, the higher the clock rate the more heat produced, this excessive heat might cause transistors damage.

CPU Utilization

HP Pavilion dm4 has the processor speed of 3 GHz which is higher than ASUS K43SJ’s 1.2 GHz. However, high processor speeds are not always better; it should also consider total memory space into comparison. For example, if you have a low processor speed with a large size of RAM, as the result, the computer will not be fully utilized. Therefore, we should balance the processor speed with total memory space to make a decision of which computer is better in terms of CPU utilization.

HP Pavilion dm4

ASUS K43SJ

<<Processor Arithmetic>>

Performance VS Power

Processor(s) Power

35.000W

45.000W

Aggregate Native Performance

2517.26MOPS/W

703.71MOPS/W

Dhrystone Integer Native

3142.54MIPS/W

895.29MIPS/W

Whetstone Double Native

2016.40MFLOPS/W

553.13MFLOPS/W

Comments:

Performance Throughput

As previously mentioned, HP Pavilion dm4 might have higher heat dissipation due to the high clock speed. However, in the power efficiency measurements, it states that HP Pavilion dm4 has higher performance throughputs per watt (power unit) over ASUS K43SJ. This measurement tells us that HP Pavilion dm4 consumes lesser power to perform the same amount of performance throughputs than ASUS K43SJ. This will result in reducing the probabilities of the excessive heat problems. In conclusion, Intel Core i7-3612QM is better because of it is not only improved in its clock speed, but at the same time it also reduces the processor power needed to perform instructions.

HP Pavilion dm4

ASUS K43SJ

<<Motherboard>>

Maximum Memory Bus Bandwidth

25GB/s

20.84GB/s

Comments:

Performance Throughput

Memory bus bandwidth (usually expressed in unit of bytes/second) refers to the rate at which data can be read from or stored into a memory by a processor. The higher bandwidth the faster the data transmission is. HP Pavilion dm4 has higher bandwidth compared to ASUS. With the faster data transmission rate, HP Pavilion dm4 is able to perform more tasks.