How does ram work?
- Thread starter Tax6132
- Start date
Solution
- May 27, 2010
- 25,816
- 4,164
- 139,340
In a broad sense that is correct. Not so much the bit about crashing, and it doesn't stop slowdowns, it just makes it harder to have them happen if you take my meaning.
Having more memory only improves performance if there is a use for it. If you aren't using 8GB of memory, then adding another 24GB won't do you much. Using more unbuffered/unregistered memory is actually statistically more likely to lead to a crash. (More chance for a bit to get misplaced)
When you start reaching the limits of your system's memory, virtual memory is an option as mentioned above. This is slower memory, so retrieval of needed data becomes slower and adds clock cycles of idle time or waiting on the CPU.
Contemporary computers don't suffer too much from a lack of memory. Aside from the slowdown to the end user's tasks, the CPU will scale back what it is doing since it has idle time waiting for data, this saves power. Much older computers tended to waste clock cycles when erroneous or unnecessary data made it into the pipeline.
Overall computers are only as fast as the data can get to the CPU or GPU. Advances in memory technologies should make the next few years interesting. 3D Xpoint from Micron/Intel is a hybrid RAM/SSD technology to help bridge the gap between SSDs and RAM.
Intel has experimented with adding DRAM caches to their CPUs onboard GPUs and allowed the CPU access to it. It had a potent effect on some tasks.
The ultimate memory is still onboard the CPU in the form of level 1 cache, this runs at the same frequency as the CPU, which is fed by slower Level 2 cache operating at the bus speed (ring I think), which is fed by the RAM, which is fed by the storage medium of the computer. All handled by predictive algorithms that anticipate what the CPU will need, when it gets it wrong you add a clock cycle.
Having more memory only improves performance if there is a use for it. If you aren't using 8GB of memory, then adding another 24GB won't do you much. Using more unbuffered/unregistered memory is actually statistically more likely to lead to a crash. (More chance for a bit to get misplaced)
When you start reaching the limits of your system's memory, virtual memory is an option as mentioned above. This is slower memory, so retrieval of needed data becomes slower and adds clock cycles of idle time or waiting on the CPU.
Contemporary computers don't suffer too much from a lack of memory. Aside from the slowdown to the end user's tasks, the CPU will scale back what it is doing since it has idle time waiting for data, this saves power. Much older computers tended to waste clock cycles when erroneous or unnecessary data made it into the pipeline.
Overall computers are only as fast as the data can get to the CPU or GPU. Advances in memory technologies should make the next few years interesting. 3D Xpoint from Micron/Intel is a hybrid RAM/SSD technology to help bridge the gap between SSDs and RAM.
Intel has experimented with adding DRAM caches to their CPUs onboard GPUs and allowed the CPU access to it. It had a potent effect on some tasks.
The ultimate memory is still onboard the CPU in the form of level 1 cache, this runs at the same frequency as the CPU, which is fed by slower Level 2 cache operating at the bus speed (ring I think), which is fed by the RAM, which is fed by the storage medium of the computer. All handled by predictive algorithms that anticipate what the CPU will need, when it gets it wrong you add a clock cycle.
TRENDING THREADS
-
-
-
Question Help in troubleshooting possible graphics card issue?
- Started by nort99
- Replies: 4
-
Question RX 6600 visual glitches right after upgrading to it- Started by Wamek
- Replies: 4
-
News US sanctions transform China into legacy chip production juggernaut — production jumped 40% in Q1 2024- Started by Admin
- Replies: 35
-
Facebook
Twitter
Instagram