L2 cache (TCC)

The L2 cache is the coherence point for current AMD Instinct™ MI-series GCN™ GPUs and CDNA™ accelerators, and is shared by all CUs on the device. Besides serving requests from the vector L1 data caches, the L2 cache also is responsible for servicing requests from the L1 instruction caches, the scalar L1 data caches and the command processor. The L2 cache is composed of a number of distinct channels (32 on MI100 and MI2XX series CDNA accelerators at 256B address interleaving) which can largely operate independently. Mapping of incoming requests to a specific L2 channel is determined by a hashing mechanism that attempts to evenly distribute requests across the L2 channels. Requests that miss in the L2 cache are passed out to Infinity Fabric™ to be routed to the appropriate memory location.

The L2 cache metrics reported by ROCm Compute Profiler are broken down into four categories:

• L2 Speed-of-Light

• L2 cache accesses

• L2-Fabric transactions

• L2-Fabric stalls

L2 Speed-of-Light

Warning

The theoretical maximum throughput for some metrics in this section are currently computed with the maximum achievable clock frequency, as reported by rocminfo, for an accelerator. This may not be realistic for all workloads.

The L2 cache’s speed-of-light table contains a few key metrics about the performance of the L2 cache, aggregated over all the L2 channels, as a comparison with the peak achievable values of those metrics:

Metric	Description	Unit
Utilization	The ratio of the number of cycles an L2 channel was active, summed over all L2 channels on the accelerator over the total L2 cycles.	Percent
Bandwidth	The number of bytes looked up in the L2 cache, as a percent of the peak theoretical bandwidth achievable on the specific accelerator. The number of bytes is calculated as the number of cache lines requested multiplied by the cache line size. This value does not consider partial requests, so e.g., if only a single value is requested in a cache line, the data movement will still be counted as a full cache line.	Percent
Hit Rate	The ratio of the number of L2 cache line requests that hit in the L2 cache over the total number of incoming cache line requests to the L2 cache.	Percent
L2-Fabric Read BW	The number of bytes read by the L2 over the Infinity Fabric interface per unit time.	GB/s
L2-Fabric Write and Atomic BW	The number of bytes sent by the L2 over the Infinity Fabric interface by write and atomic operations per unit time.	GB/s

Note

The L2 cache on AMD Instinct MI CDNA accelerators uses a “hit-on-miss” approach to reporting cache hits. That is, if while satisfying a miss, another request comes in that would hit on the same pending cache line, the subsequent request will be counted as a ‘hit’. Therefore, it is also important to consider the latency metric in the L2-Fabric section when evaluating the L2 hit rate.

L2 cache accesses

This section details the incoming requests to the L2 cache from the vL1D and other clients – for instance, the sL1D and L1I caches.

Metric	Description	Unit
Bandwidth	The number of bytes looked up in the L2 cache, per normalization unit. The number of bytes is calculated as the number of cache lines requested multiplied by the cache line size. This value does not consider partial requests, so for example, if only a single value is requested in a cache line, the data movement will still be counted as a full cache line.	Bytes per normalization unit.
Requests	The total number of incoming requests to the L2 from all clients for all request types, per normalization unit.	Requests per normalization unit.
Read Requests	The total number of read requests to the L2 from all clients.	Requests per normalization unit
Write Requests	The total number of write requests to the L2 from all clients.	Requests per normalization unit
Atomic Requests	The total number of atomic requests (with and without return) to the L2 from all clients.	Requests per normalization unit
Streaming Requests	The total number of incoming requests to the L2 that are marked as streaming. The exact meaning of this may differ depending on the targeted accelerator, however on an MI2XX this corresponds to non-temporal load or stores. The L2 cache attempts to evict streaming requests before normal requests when the L2 is at capacity.	Requests per normalization unit
Probe Requests	The number of coherence probe requests made to the L2 cache from outside the accelerator. On an MI2XX, probe requests may be generated by, for example, writes to fine-grained device memory or by writes to coarse-grained device memory.	Requests per normalization unit
Hit Rate	The ratio of the number of L2 cache line requests that hit in the L2 cache over the total number of incoming cache line requests to the L2 cache.	Percent
Hits	The total number of requests to the L2 from all clients that hit in the cache. As noted in the Speed-of-Light section, this includes hit-on-miss requests.	Requests per normalization unit
Misses	The total number of requests to the L2 from all clients that miss in the cache. As noted in the Speed-of-Light section, these do not include hit-on-miss requests.	Requests per normalization unit
Writebacks	The total number of L2 cache lines written back to memory for any reason. Write-backs may occur due to user code (such as HIP kernel calls to __threadfence_system or atomic built-ins) by the command processor’s memory acquire/release fences, or for other internal hardware reasons.	Cache lines per normalization unit
Writebacks (Internal)	The total number of L2 cache lines written back to memory for internal hardware reasons, per normalization unit.	Cache lines per normalization unit.
Writebacks (vL1D Req)	The total number of L2 cache lines written back to memory due to requests initiated by the vL1D cache, per normalization unit.	Cache lines per normalization unit.
Evictions (Normal)	The total number of L2 cache lines evicted from the cache due to capacity limits, per normalization unit.	Cache lines per normalization unit.
Evictions (vL1D Req)	The total number of L2 cache lines evicted from the cache due to invalidation requests initiated by the vL1D cache, per normalization unit.	Cache lines per normalization unit.
Non-hardware-Coherent Requests	The total number of requests to the L2 to Not-hardware-Coherent (NC) memory allocations, per normalization unit. See the Memory type for more information.	Requests per normalization unit.
Uncached Requests	The total number of requests to the L2 that go to Uncached (UC) memory allocations. See the Memory type for more information.	Requests per normalization unit.
Coherently Cached Requests	The total number of requests to the L2 that go to Coherently Cacheable (CC) memory allocations. See the Memory type for more information.	Requests per normalization unit.
Read/Write Coherent Requests	The total number of requests to the L2 that go to Read-Write coherent memory (RW) allocations. See the Memory type for more information.	Requests per normalization unit.

Note

All requests to the L2 are for a single cache line’s worth of data. The size of a cache line may vary depending on the accelerator, however on an AMD Instinct CDNA2 MI2XX accelerator, it is 128B, while on an MI100, it is 64B.

L2-Fabric transactions

Requests/data that miss in the L2 must be routed to memory in order to service them. The backing memory for a request may be local to this accelerator (i.e., in the local high-bandwidth memory), in a remote accelerator’s memory, or even in the CPU’s memory. Infinity Fabric is responsible for routing these memory requests/data to the correct location and returning any fetched data to the L2 cache. The Request flow describes the flow of these requests through Infinity Fabric in more detail, as described by ROCm Compute Profiler metrics, while Metrics give detailed definitions of individual metrics.

Request flow

The following is a diagram that illustrates how L2↔Fabric requests are reported by ROCm Compute Profiler:

Fig. 47 L2↔Fabric transaction flow on AMD Instinct MI-series accelerators.

Requests from the L2 Cache are broken down into two major categories, read requests and write requests (at this granularity, atomic requests are treated as writes).

From there, these requests can additionally subdivided in a number of ways. First, these requests may be sent across Infinity Fabric as different transaction sizes, 32B or 64B on current CDNA accelerators.

Note

On current CDNA accelerators, the 32B read request path is expected to be unused and so is disconnected in the flow diagram.

In addition, the read and write requests can be further categorized as:

• Uncached read/write requests, for instance: for access to fine-grained memory

• Atomic requests, for instance: for atomic updates to fine-grained memory

• HBM read/write requests OR remote read/write requests, for instance: for requests to the accelerator’s local HBM OR requests to a remote accelerator’s HBM or the CPU’s DRAM

These classifications are not necessarily exclusive. For example, a write request can be classified as an atomic request to the accelerator’s local HBM, and an uncached write request. The request-flow diagram marks exclusive classifications as a splitting of the flow, while non-exclusive requests do not split the flow line. For example, a request is either a 32B Write Request OR a 64B Write request, as the flow splits at this point:

Fig. 48 Splitting request flow

However, continuing along, the same request might be an atomic request and an uncached write request, as reflected by a non-split flow:

Fig. 49 Non-splitting request flow

Finally, we note that uncached read requests (e.g., to fine-grained memory) are handled specially on CDNA accelerators, as indicated in the request flow diagram. These are expected to be counted as a 64B Read Request, and if they are requests to uncached memory (denoted by the dashed line), they will also be counted as two uncached read requests (that is, the request is split):

Fig. 50 Uncached read-request splitting.

Metrics

The following metrics are reported for the L2-Fabric interface:

Metric	Description	Unit
L2-Fabric Read Bandwidth	The total number of bytes read by the L2 cache from Infinity Fabric per normalization unit.	Bytes per normalization unit.
HBM Read Traffic	The percent of read requests generated by the L2 cache that are routed to the accelerator’s local high-bandwidth memory (HBM). This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric Read bandwidth directed to the local HBM.	Percent
Remote Read Traffic	The percent of read requests generated by the L2 cache that are routed to any memory location other than the accelerator’s local high-bandwidth memory (HBM) – for example, the CPU’s DRAM or a remote accelerator’s HBM. This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric Read bandwidth directed to a remote location.	Percent
Uncached Read Traffic	The percent of read requests generated by the L2 cache that are reading from an uncached memory allocation. Note, as described in the request flow section, a single 64B read request is typically counted as two uncached read requests. So, it is possible for the Uncached Read Traffic to reach up to 200% of the total number of read requests. This breakdown does not consider the size of the request (i.e., 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric read bandwidth directed to an uncached memory location.	Percent
L2-Fabric Write and Atomic Bandwidth	The total number of bytes written by the L2 over Infinity Fabric by write and atomic operations per normalization unit. Note that on current CDNA accelerators, such as the MI2XX, requests are only considered atomic by Infinity Fabric if they are targeted at non-write-cacheable memory, for example, fine-grained memory allocations or uncached memory allocations on the MI2XX.	Bytes per normalization unit.
HBM Write and Atomic Traffic	The percent of write and atomic requests generated by the L2 cache that are routed to the accelerator’s local high-bandwidth memory (HBM). This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric Write and Atomic bandwidth directed to the local HBM. Note that on current CDNA accelerators, such as the MI2XX, requests are only considered atomic by Infinity Fabric if they are targeted at fine-grained memory allocations or uncached memory allocations.	Percent
Remote Write and Atomic Traffic	The percent of read requests generated by the L2 cache that are routed to any memory location other than the accelerator’s local high-bandwidth memory (HBM) – for example, the CPU’s DRAM or a remote accelerator’s HBM. This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric Read bandwidth directed to a remote location. Note that on current CDNA accelerators, such as the MI2XX, requests are only considered atomic by Infinity Fabric if they are targeted at fine-grained memory allocations or uncached memory allocations.	Percent
Atomic Traffic	The percent of write requests generated by the L2 cache that are atomic requests to any memory location. This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric Read bandwidth directed to a remote location. Note that on current CDNA accelerators, such as the MI2XX, requests are only considered atomic by Infinity Fabric if they are targeted at fine-grained memory allocations or uncached memory allocations.	Percent
Uncached Write and Atomic Traffic	The percent of write and atomic requests generated by the L2 cache that are targeting uncached memory allocations. This breakdown does not consider the size of the request (meaning that 32B and 64B requests are both counted as a single request), so this metric only approximates the percent of the L2-Fabric read bandwidth directed to uncached memory allocations.	Percent
Read Latency	The time-averaged number of cycles read requests spent in Infinity Fabric before data was returned to the L2.	Cycles
Write Latency	The time-averaged number of cycles write requests spent in Infinity Fabric before a completion acknowledgement was returned to the L2.	Cycles
Atomic Latency	The time-averaged number of cycles atomic requests spent in Infinity Fabric before a completion acknowledgement (atomic without return value) or data (atomic with return value) was returned to the L2.	Cycles
Read Stall	The ratio of the total number of cycles the L2-Fabric interface was stalled on a read request to any destination (local HBM, remote PCIe® connected accelerator or CPU, or remote Infinity Fabric connected accelerator [1] or CPU) over the total active L2 cycles.	Percent
Write Stall	The ratio of the total number of cycles the L2-Fabric interface was stalled on a write or atomic request to any destination (local HBM, remote accelerator or CPU, PCIe connected accelerator or CPU, or remote Infinity Fabric connected accelerator [1] or CPU) over the total active L2 cycles.	Percent

Detailed transaction metrics

The following metrics are available in the detailed L2-Fabric transaction breakdown table:

Metric	Description	Unit
32B Read Requests	The total number of L2 requests to Infinity Fabric to read 32B of data from any memory location, per normalization unit. See Request flow for more detail. Typically unused on CDNA accelerators.	Requests per normalization unit.
Uncached Read Requests	The total number of L2 requests to Infinity Fabric to read uncached data from any memory location, per normalization unit. 64B requests for uncached data are counted as two 32B uncached data requests. See Request flow for more detail.	Requests per normalization unit.
64B Read Requests	The total number of L2 requests to Infinity Fabric to read 64B of data from any memory location, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
HBM Read Requests	The total number of L2 requests to Infinity Fabric to read 32B or 64B of data from the accelerator’s local HBM, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
Remote Read Requests	The total number of L2 requests to Infinity Fabric to read 32B or 64B of data from any source other than the accelerator’s local HBM, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
32B Write and Atomic Requests	The total number of L2 requests to Infinity Fabric to write or atomically update 32B of data to any memory location, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
Uncached Write and Atomic Requests	The total number of L2 requests to Infinity Fabric to write or atomically update 32B or 64B of uncached data, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
64B Write and Atomic Requests	The total number of L2 requests to Infinity Fabric to write or atomically update 64B of data in any memory location, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
HBM Write and Atomic Requests	The total number of L2 requests to Infinity Fabric to write or atomically update 32B or 64B of data in the accelerator’s local HBM, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
Remote Write and Atomic Requests	The total number of L2 requests to Infinity Fabric to write or atomically update 32B or 64B of data in any memory location other than the accelerator’s local HBM, per normalization unit. See Request flow for more detail.	Requests per normalization unit.
Atomic Requests	The total number of L2 requests to Infinity Fabric to atomically update 32B or 64B of data in any memory location, per normalization unit. See Request flow for more detail. Note that on current CDNA accelerators, such as the MI2XX, requests are only considered atomic by Infinity Fabric if they are targeted at non-write-cacheable memory, such as fine-grained memory allocations or uncached memory allocations on the MI2XX.	Requests per normalization unit.

L2-Fabric interface stalls

When the interface between the L2 cache and Infinity Fabric becomes backed up by requests, it may stall, preventing the L2 from issuing additional requests to Infinity Fabric until prior requests complete. This section gives a breakdown of what types of requests in a kernel caused a stall (like read versus write), and to which locations – for instance, to the accelerator’s local memory, or to remote accelerators or CPUs.

Metric	Description	Unit
Read - PCIe Stall	The number of cycles the L2-Fabric interface was stalled on read requests to remote PCIe connected accelerators [1] or CPUs as a percent of the total active L2 cycles.	Percent
Read - Infinity Fabric Stall	The number of cycles the L2-Fabric interface was stalled on read requests to remote Infinity Fabric connected accelerators [1] or CPUs as a percent of the total active L2 cycles.	Percent
Read - HBM Stall	The number of cycles the L2-Fabric interface was stalled on read requests to the accelerator’s local HBM as a percent of the total active L2 cycles.	Percent
Write - PCIe Stall	The number of cycles the L2-Fabric interface was stalled on write or atomic requests to remote PCIe connected accelerators [1] or CPUs as a percent of the total active L2 cycles.	Percent
Write - Infinity Fabric Stall	The number of cycles the L2-Fabric interface was stalled on write or atomic requests to remote Infinity Fabric connected accelerators [1] or CPUs as a percent of the total active L2 cycles.	Percent
Write - HBM Stall	The number of cycles the L2-Fabric interface was stalled on write or atomic requests to accelerator’s local HBM as a percent of the total active L2 cycles.	Percent
Write - Credit Starvation	The number of cycles the L2-Fabric interface was stalled on write or atomic requests to any memory location because too many write/atomic requests were currently in flight, as a percent of the total active L2 cycles.	Percent

Warning

On current CDNA accelerators and GCN GPUs, these L2↔Fabric stalls can be undercounted in some circumstances.

Footnotes

[1] (1,2,3,4,5,6)

In addition to being used for on-accelerator data-traffic, AMD Infinity Fabric technology can be used to connect multiple accelerators to achieve advanced peer-to-peer connectivity and enhanced bandwidths over traditional PCIe connections. Some AMD Instinct MI-series accelerators like the MI250X feature coherent CPU↔accelerator connections built using AMD Infinity Fabric.

Disclaimer

PCIe® is a registered trademark of PCI-SIG Corporation.