The purpose of this document is to guide users through overclocking NVIDIA
nForce 590 SLI technology chipset and use of the numerous tools and features that
will boost system performance, enhancing end-user gaming experience.
Overclocking PCs was once quite difficult and required people to physically change
hardware. In the past, companies that designed motherboards, CPUs, memory, and
graphics processor units were not supportive of overclocking nor were they
supportive of allowing users to tamper with qualified and tested settings.
Things have changed recently, however. Companies are more accommodating of
the urgent requests PC enthusiasts make. In fact, ASIC designers now incorporate a
lot of performance and overclocking headroom in the chips they produce.
Nowadays, with the availability of highly overclockable PC components and the
utilities that allow users to easily attain higher performance, overclocking has moved
into the mainstream.
NVIDIA
overclockable. These products have provided higher performance for PC gaming
enthusiasts searching for the ultimate gaming experience.
NVIDIA specifically designed the NVIDIA nForce
based on the new AMD AM2 Athlon processor, with overclocking enthusiasts in
mind. Even though the chipset itself is highly overclockable, NVIDIA built an
entire support platform around the NVIDIA nForce 590 SLI core logic. This
platform includes:
®
GPUs have enjoyed good reputations as products that are highly
®
590 SLI™ technology chipset,
Software utilities such as the newly redesigned NVIDIA nTune 5.0 and its
highly sophisticated SBIOS
Special high-performance memory with Enhanced Performance Profiles
Overclocking refers to operating PC components beyond what the manufacturer
guarantees. When overclocking is not carried on with care, it may lead to system
instability, crashes and loss of data, and may even destroy components. However,
when overclocking is performed attentively, it increases overall system performance
quite dramatically.
Cooling
Cooling PC components sufficiently is critical to overclocking. Keeping the CPU,
GPU, or overall system cool all contribute to overall system overclockability.
Otherwise, higher component temperatures lead to operation instability and
increased power consumption.
Theoretically, frequency is directly proportional to power and temperature. The
higher the frequency, the more the power consumed and the hotter the device
becomes. In order to achieve higher frequencies, component temperatures should
be kept as low as possible. With proper cooling, overclocking is highly attainable.
The most commonly cooling techniques are air cooling with fan-mounted heat sinks
and water cooling blocks. Even though water cooling is harder to install, it achieves
higher overclockability in comparison to air cooling because it keeps temperatures
lower. With technological advancements, air cooling has improved in performance
providing the means for overclocking. Even so, water cooling CPUs and GPUs is
becoming more widely accepted.
For NVIDIA recommended CPU coolers, visit
components.
Power Supply Units
Power supply units are also important when it comes to overclocking. Increasing
system performance leads to increased power consumption by the overclocked
components. Having a power supply that generates enough power with headroom
is essential in guaranteeing system stability. Check the following site for a list of
power supplies suitable for SLI systems:
Synthetic 3D benchmark that provides a measure of graphics
performance. It is also often used to test a system for stability.
Technology that changes CPU frequency dynamically depending
™
on power and thermal requirements of a CPU. Because it limits
the overclockability of an AMD CPU, it needs to be turned off
for best results.
For more information, visit this link
HyperTransport interface connecting the AMD CPU to NVIDIA
nForce 590 SLI SPP. Nominally, it runs at 1 GHz clock or 4
GBps in downstream and upstream directions.
Clock signal of the CPU HyperTransport link. It has impact on
HyperTransport bus speed, CPU frequency, as well as memory
frequency. It is 200 MHz nominally.
Multiplier for your front-side bus (FSB). Multiply this value by the
speed of your FSB to get the end CPU clock speed:
CPU Clock Speed = Multiplier x FSB
Program that searches for very large prime number. It contains
stability tests and is an indication of how robust a system is.
Serious Sam
Video game which is used by overclockers to test system
robustness.
t
RAS
Row Address Strobe time is the amount of time between a row
being activated by precharge and deactivated. A row cannot be
deactivated until
has completed. The lower t
RAS
is, the faster
RAS
t
the performance. If it is set too low, it can cause data corruption
by deactivating the row too soon.
tWR
Write Recovery time is the memory timing that determines the
delay between a write command and when a precharge command
is set to the same bank of memory.
t
WRRD
W to R Termination Turnaround time is the number of clock
cycles between the last write data pair and the subsequent READ
command to the same physical bank.
t
RCD
RAS to CAS Access time is the amount of time in cycles for
issuing an active command and the read/write commands.
RAS to RAS Delay is the amount of cycles that it takes to activate
the next bank of memory. The lower the timing, the better the
performance but it can cause instability.
Refresh Rate is the amount of time it takes before a change is
refreshed. If the charge is not refreshed enough, the signal loses
its charge and corrupts data. It is measured in microseconds.
Row Precharge time is the minimum time between active
commands and the read/writes of the next bank of the memory
module.
Read to Write Delay is when a write command is received, this is
the amount of cycles for the command to be executed.
Read to Read Timing is the number of clock cycles between the
last read and the subsequent READ command to the same
physical bank.
Row Cycle time is the minimum time in cycles it takes a row to
t
= t
+ t
complete a full cycle. This can be determined by
t
is set too short, it can cause data corruption. If t
If
RC
RC
RC
RAS
is set too
.
RP
long, stability increases at the expense of performance.
Write to Read Delay is the amount of cycles required between a
valid write command and the next read command. Lower is
better performance but can cause instability.
t
tCL
t
WRWR
CPC
Write to Write time is the number of clock cycles between the last
write and the subsequent WRITE command to the same physical
bank.
CAS latency is the number of clock cycles between the memory
receiving a
READ command and actually starting to read
Command per Clock has only two values: 1 and 2. When set to 1,
it provides the best performance but limits memory frequency. In
order to overclock memory frequency, a value of 2 is more
reasonable.
DU-02451-001_v01 4
05/17/06
Overclocking the NVIDIA
nForce 590 SLI Technology Chipset
This section describes how to maximize the performance of an NVIDIA nForce
590 SLI PC system. It provides a systematic approach to unleashing its immense
performance potential. Bus speeds, bandwidth and processor core speeds — both
CPU and GPU core speeds — are the main contributing factors to boosting overall
system performance.
This section describes how to adjust settings for the major chip interconnects and
components that affect performance. Figure 1 shows a high-level diagram of
components in a PC system:
AMD Athlon CPU core frequency
DDR2 memory interface bandwidth and latency
CPU-NVIDIA nForce 590 SLI SPP HyperTransport bus
NVIDIA nForce 590 SLI SPP-NVIDIA nForce 590 SLI MCP
HyperTransport bus
NVIDIA nForce 590 SLI SPP/MCP PCI Express bus
GPU core and memory frequency
Before overclocking, disable the AMD Cool’n’Quiet™ technology.
The AMD Cool’n’Quiet technology must be disabled in order to enable stable
overclocking without limiting maximum achieved CPU performance. Use the
following procedure to disable this feature:
1. Set AMD Cool’n’Quiet to Disable on the Advanced Power Management
A step-by-step approach helps simplify the complex task of overclocking, described
in the following steps:
1. CPU reference clock (HTT)
2. CPU HyperTransport interface speed
3. CPU core clock frequency
4. DDR2 memory interface frequency & timing
5. Balance the results achieved with CPU frequency, HyperTransport, and
memory overclocking
6. Overclock the PCI Express x16 buses
7. Increase the NVIDIA nForce 590 SLI MCP HyperTransport bus frequency
8. Boost the GPU core and memory frequencies
Each of the following sections provides guidance on how to tweak an NVIDIA
nForce 590-SLI technology chipset based PC system based on this step-by-step
approach.
The first steps for overclocking the CPU reference clock, CPU HyperTransport bus,
CPU frequency, memory interface, PCI Express, and MCP HyperTransport link are
done in the SBIOS. Some system settings, such as CPU multiplier and memory
drive strength parameters, can only be adjusted in the SBIOS.
Once you use the SBIOS for an initial pass at overclocking, you then fine-tune the
settings with NVIDIA nTune 5.0. NVIDIA nTune provides flexibility to adjust
majority of parameters that affect overclocking in a Windows environment, allowing
for dynamic system overclocking and stability testing without the need for constant
system reboots.