GAP Message (Diagnostic ID 30756)

Intel® C++ Compiler Classic Developer Guide and Reference

Download PDF

ID 767249

Date 3/31/2023

Version

Public

A newer version of this document is available. Customers should click here to go to the newest version.

Visible to Intel only — GUID: GUID-179A7895-223F-4A3C-8E25-941FCF6C8F99

View Details

GAP Message (Diagnostic ID 30756)

Message

Split the structure '%s' into two parts to improve data locality. Frequently accessed fields are '%s'; performance may improve by putting these fields into one structure and the remaining fields into another structure. Alternatively, performance may also improve by reordering the fields of the structure. Suggested field order: '%s'.

Advice

This message is issued when both structure splitting and field reordering transformations are applicable. Structure splitting transformation is expected to lead to higher performance gains if the transformation can be successfully applied. However, field reordering transformation is usually simple enough to apply, but the downside is that the performance gain seen may be lower.

You must verify that the structure meets the requirements for applying the splitting or reordering transformation. Some of these requirements are described in the description of these individual transformations.

Example

Consider the following:

//str_split_reord.c 
struct str {
    int a1, b1, carr[100], c1, e1; 
};
 
#define N 1000000
 
struct str *sp;
 
void allocate_str_mem() {
    sp = malloc(N * sizeof(struct str)); 
}
 
int hot_func1() {
    int i, ret = 0;
 
    for (i = 0; i < 1000000; i++) {
        ret += sp[i].a1;
        ret += sp[i].c1;
    }
    sp->carr[0] = ret;
    return ret; 
}
 
int hot_func2() {
    int ret = 0, i;
    for (i = 0; i < 100000; i++) {
        ret += sp[i].a1;
        ret -= sp[i].e1;
    }
    return ret; 
}
 
int hot_func3() {
    int ret = 0, i;
    for (i = 0; i < 1000000; i++) {
        ret += sp[i].b1;
    }
    return ret; 
}

In this case, a message is displayed that is similar to the following:

drive: program-name: remark #30756: (DTRANS) Split the structure 'str' into two parts to improve data locality. Frequently accessed fields are 'a1, b1, c1'; performance may improve by putting these fields into one structure and the remaining fields into another structure. Alternatively, performance may also improve by reordering the fields of the structure. Suggested field order: 'a1, c1, e1, b1, carr'. ...(etc.)

If you determine it is safe to do so, you can modify the program code as shown below to split the structure 'str'. Other references to structure 'str' that are not in the current module should also be similarly modified.

struct str_cold {
    int carr[100], e1; 
};
 
struct str {
    int a1, b1, c1; struct str_cold *cold_ptr; 
};
 
#define N 1000000
 
struct str *sp;
 
void allocate_str_mem() 
{
    struct str *temp;
    struct str_cold *cold_begin;
    int index;
 
    temp = malloc(N * sizeof(struct str) + N * sizeof(struct str_cold));
    sp = temp;
    cold_begin = (struct str_cold *) (temp + N);
    for(index = 0; index < N; index++) {
       temp[index].cold_ptr = cold_begin + index;
    } 
}
 
int hot_func1() {
    int i, ret = 0;
 
    for (i = 0; i < 1000000; i++) {
        ret += sp[i].a1;
        ret += sp[i].c1;
    }
    sp->cold_ptr->carr[0] = ret;
    return ret; 
}
 
int hot_func2() {
    int ret = 0, i;
    for (i = 0; i < 100000; i++) {
        ret += sp[i].a1
        ret -= sp[i].cold_ptr->e1;
    }
    return ret; 
}
 
int hot_func3() {
    int ret = 0, i;
    for (i = 0; i < 1000000; i++) {
        ret += sp[i].b1;
    }
    return ret; 
}

For the above example, the only source change required to reorder fields in structure 'str' as alternatively suggested are the following:

//str_split_reord.c 
struct str {
    int a1, c1, e1, b1, carr[100]; 
}; 
...

Verify

The suggestion is based on the field references in the current compilation. Please make sure that the restructuring is applied to field references in all source files of the application, and that the restructured code satisfies the original program semantics.

Parent topic: Guided Auto Parallelism Messages

Select Your Language

Using Intel.com Search

Quick Links

Recent Searches

Advanced Search

Only search in

Intel® C++ Compiler Classic Developer Guide and Reference

GAP Message (Diagnostic ID 30756)

Message

Advice

Example

Verify