Thermal constraints can be significant in Flash storage systems. One way to mitigate temperature issues is by using a heat sink. Another way is to throttle the performance of the storage system. Throttling is often used when a heat sink is not a viable option for the form factor of the storage system. Some storage systems employ static throttling where fixed delays are introduced when the temperature of the storage system reaches a certain threshold.
Overview
By way of introduction, the below embodiments relate to a storage system and method for data-driven intelligent thermal throttling. In one embodiment, a storage system is provided comprising a memory and a controller. The controller is configured to determine a temperature of the memory; estimate a future temperature curve based on the temperature of the memory; and determine a memory throttling delay to apply based on the estimated future temperature curve.
In some embodiments, the controller is further configured to estimate the future temperature curve based on a saturating temperature (K) and a curvature factor (R).
In some embodiments, the controller is further configured to sample the temperature of the memory over a plurality of time windows.
In some embodiments, at least one of the following dynamically changes based on the temperature of the memory: a length of a time window, a frequency of sampling the temperature of the memory in a time window, and a number of time windows in the plurality of time windows.
In some embodiments, the controller is further configured to estimate the future temperature curve based on the sampled temperature in each time window.
In some embodiments, the controller is further configured to estimate the future temperature curve by minimizing a sum of squares of errors.
In some embodiments, the controller is further configured to estimate a new future temperature curve after applying the determined memory throttling delay.
In some embodiments, the controller is further configured to apply static throttling as an exception handling operation.
In some embodiments, the memory comprises a three-dimensional memory.
In another embodiment, a method is provided that is performed in a storage system comprising a memory. The method comprises predicting a curve of temperatures that the memory will experience and dynamically throttling memory operations based on the predicted curve.
In some embodiments, the curve is predicted based on an actual temperature of the memory, a saturating temperature (K), and a curvature factor (R).
In some embodiments, the method further comprises sampling temperature of the memory over a plurality of time windows.
In some embodiments, at least one of the following dynamically changes based on the temperature of the memory: a length of a time window, a frequency of sampling the temperature of the memory in a time window, and a number of time windows in the plurality of time windows.
In some embodiments, the method further comprises using a minimizing-a-sum-of-squares-of-errors technique to predict the curve.
In some embodiments, the method further comprises predicting a new curve after dynamically throttling the memory operations.
In some embodiments, the method further comprises applying static throttling in response to an exception.
In another embodiment, a storage system is provided comprising a memory, means for predicting a future temperature of the memory, and means for throttling memory operations based on the predicted future temperature of the memory.
In some embodiments, the predicted future temperature of the memory is based on an estimated temperature curve.
In some embodiments, the temperature curve is estimated based on an actual temperature of the memory, a saturating temperature (K), and a curvature factor (R).
In some embodiments, the temperature curve is estimated using a minimizing-a-sum-of-squares-of-errors technique.
Other embodiments are possible, and each of the embodiments can be used alone or together in combination. Accordingly, various embodiments will now be described with reference to the attached drawings.
Embodiments
Storage systems suitable for use in implementing aspects of these embodiments are shown in
The controller 102 (which may be a non-volatile memory controller (e.g., a flash, resistive random-access memory (ReRAM), phase-change memory (PCM), or magnetoresistive random-access memory (MRAM) controller)) can take the form of processing circuitry, a microprocessor or processor, and a computer-readable medium that stores computer-readable program code (e.g., firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, for example. The controller 102 can be configured with hardware and/or firmware to perform the various functions described below and shown in the flow diagrams. Also, some of the components shown as being internal to the controller can also be stored external to the controller, and other components can be used. Additionally, the phrase “operatively in communication with” could mean directly in communication with or indirectly (wired or wireless) in communication with through one or more components, which may or may not be shown or described herein.
As used herein, a non-volatile memory controller is a device that manages data stored on non-volatile memory and communicates with a host, such as a computer or electronic device. A non-volatile memory controller can have various functionality in addition to the specific functionality described herein. For example, the non-volatile memory controller can format the non-volatile memory to ensure the memory is operating properly, map out bad non-volatile memory cells, and allocate spare cells to be substituted for future failed cells. Some part of the spare cells can be used to hold firmware to operate the non-volatile memory controller and implement other features. In operation, when a host needs to read data from or write data to the non-volatile memory, it can communicate with the non-volatile memory controller. If the host provides a logical address to which data is to be read/written, the non-volatile memory controller can convert the logical address received from the host to a physical address in the non-volatile memory. (Alternatively, the host can provide the physical address.) The non-volatile memory controller can also perform various memory management functions, such as, but not limited to, wear leveling (distributing writes to avoid wearing out specific blocks of memory that would otherwise be repeatedly written to) and garbage collection (after a block is full, moving only the valid pages of data to a new block, so the full block can be erased and reused).
Non-volatile memory die 104 may include any suitable non-volatile storage medium, including resistive random-access memory (ReRAM), magnetoresistive random-access memory (MRAM), phase-change memory (PCM), NAND flash memory cells and/or NOR flash memory cells. The memory cells can take the form of solid-state (e.g., flash) memory cells and can be one-time programmable, few-time programmable, or many-time programmable. The memory cells can also be single-level cells (SLC), multiple-level cells (MLC), triple-level cells (TLC), or use other memory cell level technologies, now known or later developed. Also, the memory cells can be fabricated in a two-dimensional or three-dimensional fashion.
The interface between controller 102 and non-volatile memory die 104 may be any suitable flash interface, such as Toggle Mode 200, 400, or 800. In one embodiment, storage system 100 may be a card based system, such as a secure digital (SD) or a micro secure digital (micro-SD) card. In an alternate embodiment, storage system 100 may be part of an embedded storage system.
Although, in the example illustrated in
Referring again to modules of the controller 102, a buffer manager/bus controller 114 manages buffers in random access memory (RAM) 116 and controls the internal bus arbitration of controller 102. A read only memory (ROM) 118 stores system boot code. Although illustrated in
Front-end module 108 includes a host interface 120 and a physical layer interface (PHY) 122 that provide the electrical interface with the host or next level storage controller. The choice of the type of host interface 120 can depend on the type of memory being used. Examples of host interfaces 120 include, but are not limited to, SATA, SATA Express, serially attached small computer system interface (SAS), Fibre Channel, universal serial bus (USB), PCIe, and NVMe. The host interface 120 typically facilitates transfer for data, control signals, and timing signals.
Back-end module 110 includes an error correction code (ECC) engine 124 that encodes the data bytes received from the host, and decodes and error corrects the data bytes read from the non-volatile memory. A command sequencer 126 generates command sequences, such as program and erase command sequences, to be transmitted to non-volatile memory die 104. A RAID (Redundant Array of Independent Drives) module 128 manages generation of RAID parity and recovery of failed data. The RAID parity may be used as an additional level of integrity protection for the data being written into the memory device 104. In some cases, the RAID module 128 may be a part of the ECC engine 124. A memory interface 130 provides the command sequences to non-volatile memory die 104 and receives status information from non-volatile memory die 104. In one embodiment, memory interface 130 may be a double data rate (DDR) interface, such as a Toggle Mode 200, 400, or 800 interface. A flash control layer 132 controls the overall operation of back-end module 110.
The storage system 100 also includes other discrete components 140, such as external electrical interfaces, external RAM, resistors, capacitors, or other components that may interface with controller 102. In alternative embodiments, one or more of the physical layer interface 122, RAID module 128, media management layer 138 and buffer management/bus controller 114 are optional components that are not necessary in the controller 102.
Returning again to
The FTL may include a logical-to-physical address (L2P) map and allotted cache memory. In this way, the FTL translates logical block addresses (“LBAs”) from the host to physical addresses in the memory 104. The FTL can include other features, such as, but not limited to, power-off recovery (so that the data structures of the FTL can be recovered in the event of a sudden power loss) and wear leveling (so that the wear across memory blocks is even to prevent certain blocks from excessive wear, which would result in a greater chance of failure).
Turning again to the drawings,
As mentioned above, thermal constraints can be significant in Flash storage systems. One way to mitigate temperature issues is to use a heat sink, and the other way is to throttle the performance of the storage system. Throttling is often used when a heat sink is not a viable option for the form factor of the storage system. Some storage systems employ static throttling where fixed delays are introduced when the temperature of the storage system reaches a certain threshold.
Static throttling is based on static temperature limits and throttle amounts set by engineers based on trial-and-error analysis of the temperature curve. The analysis can be done after conducting various experiments under different host/environment conditions where various static delays are introduced at various temperature thresholds. Because the thresholds and delays are fixed, they are decided to accommodate severe environment /host conditions. Because these temperature thresholds and delays are pre-defined to accommodate worst-case scenarios, a lot of performance can be wasted in most of the general cases, as the introduced delays might be far greater than required.
For example, consider the two different cases shown in
The following embodiments explain how the controller 102 of the storage system 100 can use temperature sensor data to estimate the future temperature curve and dynamically introduce delays so that the end user experiences is optimal, and sustained and smooth performance can be met under various scenarios pertaining to changes in environment, host conditions, NAND conditions, data noise, etc. This dynamic throttling can even be extended to other systems with thermal constraints, such as laptops, smartphones, etc.
More specifically, in one embodiment, to overcome the above-mentioned limitations, the controller 102 of the storage system 100 can use a data-driven intelligent thermal throttling method, where the controller 102 (e.g., firmware) utilizes memory array (e.g., NAND) 104 temperature sensor data to estimate a future temperature curve and give optimal delays at certain temperature thresholds based on the estimated curve. The throttling of memory operations can throttle read and/or write operations to the memory (e.g., reducing the number of such operations that are performed over time). Before turning to the details on this method, an equation will be derived to understand the behavior of temperature at various scenarios. This equation can be used in the curve estimation. It should be noted that this particular equation is just one possible implementation, and other implementations can be used. Also, while these examples will be discussed in terms of a storage system (e.g., in this example, a microSD card), it should be noted that the below derivation can be used with any suitable system with thermal constraints (e.g., a smartphone, laptop, etc.).
The main reason for an increase in the temperature of the storage system 100 is the power generation by the card, which depends on NAND performance. Based on the characteristics of the card and its surroundings, a part of the heat generated is dissipated to the surroundings, and the remaining part is consumed by the card to raise its temperature.
So, consider a small instance of time ‘dt’, during which:
1. ‘dG’ heat is generated by the card due to the NAND performance
2. ‘dQ’ heat is used by the card to raise the temperature by ‘dT’
3. ‘dE’ heat is released outside
Power generation by the card (P) is the rate of heat generation.
P is dependent on the NAND performance but is independent of time or temperature. And say ‘C’ is the cumulative heat capacity of the card. By definition of heat capacity,
Heat transfer to the surroundings is happening through the air by the mode of convective heat transfer. The equation of which is described as follows:
Where H is the heat transfer coefficient, A is the area available for heat transfer Ts is the surrounding temperature
From the equations 1, 2, 3, and 4:
Consider H.A=h
Ignoring the mutual dependence of K and R over the factor h for simpler estimation;
T=K−e
−Rt(K−T0)
The above equation coincides with Newton's law of cooling, which is T=Tenv−e−Rt(Tenv−T0). The fundamental difference lies in the proof, where there is no external factor like a power generation component in Newton's law, which affects the saturating temperature.
Though various parameters are assumed constant with time while deriving the equation, they can slightly vary due to various surrounding factors. However, a cumulative value of a parameter for a time range [0→t] will always hold valid for the equation.
Hence, temperature of any product can be estimated just by estimating the two parameters K and R, irrespective of change in host, device, or surrounding conditions for any product.
The below section explains the behavior of the curve with the change in K, R values.
As can be seen in the equation, T=K at t=∞. So, K is the saturating temperature.
As
K mostly depends on the surrounding temperature and the performance. As we cannot control the surrounding temperature, we can decrease performance to decrease saturating temperature.
And the other factor is ‘R,’ which decides the curvature of the temperature and how fast the temperature can reach closer to the saturating temperature.
As
R mainly depends on the device characteristics, so R does not vary much for a system. The surroundings can have an impact on the value of R. So, if the host containing the card is placed in an enclosed area (e.g., a bag), that affects the availability of airflow and surface area for heat release.
Having a heat sink would reduce the value of the ‘R,’ which makes the curve less sharp, and the rate of temperature increment is reduced. But a storage device with a small form factor has no scope to control ‘R.’
Now that the above paragraphs have explained the behavior of the curve at various scenarios, the following paragraphs show the experimental data and how well the theoretical equation matches the experimental data. These experiments were performed on a one terabyte (TB) microSD card as reference.
The curves in
In one embodiment, the controller 102 (e.g., firmware) uses temperature sensor samples to estimate the curve. The temperature sensor can be located in the memory 104, near the memory 104 within the enclosure of the storage system 100, or in any other suitable location. The temperature is sampled for a period of time (called a window), and the window can move by any number of samples each time. This is shown in
As shown in
Assume the parameter ‘R’ is known.
T=K−e
−Rt(K−T0),
Substituting K−T0=F=>K=F+T0
=>T=F+T0−eRt(F)
=>T−T0=F−eRt(F)
=>Y=F−F, e−Rt, where Y=T−T0
Error of estimation of any sample point ‘i’
E
i
=Y
i
−F+F, e
−Rt
=>Ei2=(Yi−F+F, e−Rt
=>Ei2=Yi2+F2+F2e−2Rt
Sum of squares of errors of all the sample points:
As the sum of squares of errors is minimum at the desired ‘F’:
For every given ‘R,’ the parameter ‘K’ is estimated using the above equation. The problem is to estimate ‘R.’ Varying ‘R’ on an exponential scale between a selected range and plotting error versus R after estimating ‘K’ for a window gives a V-shape curve as shown in
It is important to note that values are chosen on an exponential scale for ‘R’ where the difference between any consecutive values is proportional to the value. Because a linear scale would give a sharp needle-like curve in a big range, a lot of unnecessary samples are needed in a linear scale.
Now that the curve is estimated, the decision making for the throttling is subjective and depends on the product. For example, the model may only trigger after a certain temperature is hit (e.g., 35° C.), and the window can move by ‘n’ samples every time. If ‘R’ for a particular application does not change much, R can be estimated for every four windows using same ‘R’ every four windows, while estimating K for every window. Also, the decision making by the controller 102 (e.g., firmware) can be after every window or after fixed set temperature thresholds. In addition, to handle worst-case exceptions, a few temperature thresholds or any unexpected events can be used to trigger fixed delays. Throttling can be done in a few stages to have a smooth performance shift. In case of abrupt throttling, the window may need to start again because the value of K may have changed too much. Hardware accelerators can also be employed to reduce the delay, in case an application involves a lot of computations and thresholds.
There are several advantages associated with these embodiments. For example, these embodiments can provide increased average performance during large data transfers, dynamic decision making based on a curve prediction, a single solution to estimate the curve for various products, performance sustainment, and a smooth performance shift, such that a user would not detect any sudden throttle. Also, these embodiments can provide improved performance at high temperature conditions. As a result, instead of throttling due to NAND limitation, the storage system of these embodiments provide improved performance.
Finally, as mentioned above, any suitable type of memory can be used. Semiconductor memory devices include volatile memory devices, such as dynamic random access memory (“DRAM”) or static random access memory (“SRAM”) devices, non-volatile memory devices, such as resistive random access memory (“ReRAM”), electrically erasable programmable read only memory (“EEPROM”), flash memory (which can also be considered a subset of EEPROM), ferroelectric random access memory (“FRAM”), and magnetoresistive random access memory (“MRAM”), and other semiconductor elements capable of storing information. Each type of memory device may have different configurations. For example, flash memory devices may be configured in a NAND or a NOR configuration.
The memory devices can be formed from passive and/or active elements, in any combinations. By way of non-limiting example, passive semiconductor memory elements include ReRAM device elements, which in some embodiments include a resistivity switching storage element, such as an anti-fuse, phase change material, etc., and optionally a steering element, such as a diode, etc. Further by way of non-limiting example, active semiconductor memory elements include EEPROM and flash memory device elements, which in some embodiments include elements containing a charge storage region, such as a floating gate, conductive nanoparticles, or a charge storage dielectric material.
Multiple memory elements may be configured so that they are connected in series or so that each element is individually accessible. By way of non-limiting example, flash memory devices in a NAND configuration (NAND memory) typically contain memory elements connected in series. A NAND memory array may be configured so that the array is composed of multiple strings of memory in which a string is composed of multiple memory elements sharing a single bit line and accessed as a group. Alternatively, memory elements may be configured so that each element is individually accessible, e.g., a NOR memory array. NAND and NOR memory configurations are examples, and memory elements may be otherwise configured.
The semiconductor memory elements located within and/or over a substrate may be arranged in two or three dimensions, such as a two-dimensional memory structure or a three-dimensional memory structure.
In a two-dimensional memory structure, the semiconductor memory elements are arranged in a single plane or a single memory device level. Typically, in a two-dimensional memory structure, memory elements are arranged in a plane (e.g., in an x-z direction plane) which extends substantially parallel to a major surface of a substrate that supports the memory elements. The substrate may be a wafer over or in which the layer of the memory elements are formed or it may be a carrier substrate which is attached to the memory elements after they are formed. As a non-limiting example, the substrate may include a semiconductor such as silicon.
The memory elements may be arranged in the single memory device level in an ordered array, such as in a plurality of rows and/or columns. However, the memory elements may be arrayed in non-regular or non-orthogonal configurations. The memory elements may each have two or more electrodes or contact lines, such as bit lines and wordlines.
A three-dimensional memory array is arranged so that memory elements occupy multiple planes or multiple memory device levels, thereby forming a structure in three dimensions (i.e., in the x, y and z directions, where the y direction is substantially perpendicular and the x and z directions are substantially parallel to the major surface of the substrate).
As a non-limiting example, a three-dimensional memory structure may be vertically arranged as a stack of multiple two-dimensional memory device levels. As another non-limiting example, a three-dimensional memory array may be arranged as multiple vertical columns (e.g., columns extending substantially perpendicular to the major surface of the substrate, i.e., in the y direction) with each column having multiple memory elements in each column. The columns may be arranged in a two-dimensional configuration, e.g., in an x-z plane, resulting in a three-dimensional arrangement of memory elements with elements on multiple vertically stacked memory planes. Other configurations of memory elements in three dimensions can also constitute a three-dimensional memory array.
By way of non-limiting example, in a three-dimensional NAND memory array, the memory elements may be coupled together to form a NAND string within a single horizontal (e.g., x-z) memory device levels. Alternatively, the memory elements may be coupled together to form a vertical NAND string that traverses across multiple horizontal memory device levels. Other three-dimensional configurations can be envisioned wherein some NAND strings contain memory elements in a single memory level while other strings contain memory elements which span through multiple memory levels. Three-dimensional memory arrays may also be designed in a NOR configuration and in a ReRAM configuration.
Typically, in a monolithic three-dimensional memory array, one or more memory device levels are formed above a single substrate. Optionally, the monolithic three-dimensional memory array may also have one or more memory layers at least partially within the single substrate. As a non-limiting example, the substrate may include a semiconductor such as silicon. In a monolithic three-dimensional array, the layers constituting each memory device level of the array are typically formed on the layers of the underlying memory device levels of the array. However, layers of adjacent memory device levels of a monolithic three-dimensional memory array may be shared or have intervening layers between memory device levels.
Then again, two-dimensional arrays may be formed separately and then packaged together to form a non-monolithic memory device having multiple layers of memory. For example, non-monolithic stacked memories can be constructed by forming memory levels on separate substrates and then stacking the memory levels atop each other. The substrates may be thinned or removed from the memory device levels before stacking, but as the memory device levels are initially formed over separate substrates, the resulting memory arrays are not monolithic three-dimensional memory arrays. Further, multiple two-dimensional memory arrays or three-dimensional memory arrays (monolithic or non-monolithic) may be formed on separate chips and then packaged together to form a stacked-chip memory device.
Associated circuitry is typically required for operation of the memory elements and for communication with the memory elements. As non-limiting examples, memory devices may have circuitry used for controlling and driving memory elements to accomplish functions such as programming and reading. This associated circuitry may be on the same substrate as the memory elements and/or on a separate substrate. For example, a controller for memory read-write operations may be located on a separate controller chip and/or on the same substrate as the memory elements.
One of skill in the art will recognize that this invention is not limited to the two-dimensional and three-dimensional structures described but cover all relevant memory structures within the spirit and scope of the invention as described herein and as understood by one of skill in the art.
It is intended that the foregoing detailed description be understood as an illustration of selected forms that the invention can take and not as a definition of the invention. It is only the following claims, including all equivalents, that are intended to define the scope of the claimed invention. Finally, it should be noted that any aspect of any of the embodiments described herein can be used alone or in combination with one another.
This application claims the benefit of U.S. provisional patent application No. 63/114,854, filed Nov. 17, 2020, which is hereby incorporated by reference.
Number | Date | Country | |
---|---|---|---|
63114854 | Nov 2020 | US |